From caecd8622f09d9f2c06b02d783cdd5456d0fcef6 Mon Sep 17 00:00:00 2001
From: saem <saemghani+github@gmail.com>
Date: Tue, 21 Apr 2026 18:00:13 -0700
Subject: [PATCH 01/28] providers: add OpenAI compatible provider

Summary
=======

Created OpenAI compatible model and embedding providers.

Details
=====

The new provider is compatible with any OpenAI API endpoint. This
includes OpenRouter, which has been subsumed as part of this change.

Detection has been added for:

- OpenAI
- LM Studio
- ollama
- vllm

This should allow for easy local model detection and usage.
---
 README.md                                  |   9 ++
 src/config.ts                              |  44 ++++++
 src/providers/embedding/index.ts           |  31 +++-
 src/providers/embedding/openai.ts          |  19 +--
 src/providers/embedding/openrouter.ts      |  51 -------
 src/providers/index.ts                     |  53 +++++--
 src/providers/{openrouter.ts => openai.ts} |  45 +++---
 src/types.ts                               |   2 +-
 test/embedding-provider.test.ts            |  45 +++---
 test/local-providers.test.ts               | 158 +++++++++++++++++++++
 10 files changed, 337 insertions(+), 120 deletions(-)
 delete mode 100644 src/providers/embedding/openrouter.ts
 rename src/providers/{openrouter.ts => openai.ts} (57%)
 create mode 100644 test/local-providers.test.ts

diff --git a/README.md b/README.md
index 5fcd243..ee9d2b1 100644
--- a/README.md
+++ b/README.md
@@ -782,6 +782,10 @@ agentmemory auto-detects from your environment. No API key needed if you have a
 | **Claude subscription** (default) | No config needed | Uses `@anthropic-ai/claude-agent-sdk` |
 | Anthropic API | `ANTHROPIC_API_KEY` | Per-token billing |
 | MiniMax | `MINIMAX_API_KEY` | Anthropic-compatible |
+| OpenAI | `OPENAI_API_KEY` | Standard API |
+| LM Studio | `LMSTUDIO_BASE_URL` | Local OpenAI-compatible |
+| Ollama | `OLLAMA_BASE_URL` | Local OpenAI-compatible |
+| vLLM | `VLLM_BASE_URL` | Local OpenAI-compatible |
 | Gemini | `GEMINI_API_KEY` | Also enables embeddings |
 | OpenRouter | `OPENROUTER_API_KEY` | Any model |
 
@@ -794,10 +798,15 @@ Create `~/.agentmemory/.env`:
 # ANTHROPIC_API_KEY=sk-ant-...
 # GEMINI_API_KEY=...
 # OPENROUTER_API_KEY=...
+# OPENAI_API_KEY=sk-...
+# OLLAMA_MODEL=llama3
+# LMSTUDIO_BASE_URL=http://localhost:1234/v1/chat/completions
 
 # Embedding provider (auto-detected, or override)
 # EMBEDDING_PROVIDER=local
 # VOYAGE_API_KEY=...
+# OLLAMA_EMBEDDING_MODEL=nomic-embed-text
+# LMSTUDIO_EMBEDDING_BASE_URL=http://localhost:1234/v1/embeddings
 
 # Search tuning
 # BM25_WEIGHT=0.4
diff --git a/src/config.ts b/src/config.ts
index 45e5088..310f70c 100644
--- a/src/config.ts
+++ b/src/config.ts
@@ -53,6 +53,42 @@ function detectProvider(env: Record<string, string>): ProviderConfig {
     };
   }
 
+  if (env["LMSTUDIO_BASE_URL"] || env["LMSTUDIO_MODEL"]) {
+    return {
+      provider: "lmstudio",
+      model: env["LMSTUDIO_MODEL"] || "local-model",
+      maxTokens,
+      baseURL: env["LMSTUDIO_BASE_URL"],
+    };
+  }
+
+  if (env["OLLAMA_BASE_URL"] || env["OLLAMA_MODEL"]) {
+    return {
+      provider: "ollama",
+      model: env["OLLAMA_MODEL"] || "llama3",
+      maxTokens,
+      baseURL: env["OLLAMA_BASE_URL"] || "http://localhost:11434/v1/chat/completions",
+    };
+  }
+
+  if (env["VLLM_BASE_URL"] || env["VLLM_MODEL"]) {
+    return {
+      provider: "vllm",
+      model: env["VLLM_MODEL"] || "local-model",
+      maxTokens,
+      baseURL: env["VLLM_BASE_URL"],
+    };
+  }
+
+  if (env["OPENAI_API_KEY"] || env["OPENAI_BASE_URL"]) {
+    return {
+      provider: "openai",
+      model: env["OPENAI_MODEL"] || "gpt-4o",
+      maxTokens,
+      baseURL: env["OPENAI_BASE_URL"] || "https://api.openai.com/v1/chat/completions",
+    };
+  }
+
   if (env["ANTHROPIC_API_KEY"]) {
     return {
       provider: "anthropic",
@@ -152,6 +188,10 @@ export function detectEmbeddingProvider(
   if (source["VOYAGE_API_KEY"]) return "voyage";
   if (source["COHERE_API_KEY"]) return "cohere";
   if (source["OPENROUTER_API_KEY"]) return "openrouter";
+  if (source["OLLAMA_EMBEDDING_BASE_URL"] || source["OLLAMA_EMBEDDING_MODEL"])
+    return "ollama";
+  if (source["LMSTUDIO_EMBEDDING_BASE_URL"] || source["LMSTUDIO_EMBEDDING_MODEL"])
+    return "lmstudio";
   return null;
 }
 
@@ -254,6 +294,10 @@ const VALID_PROVIDERS = new Set([
   "openrouter",
   "agent-sdk",
   "minimax",
+  "lmstudio",
+  "openai",
+  "ollama",
+  "vllm",
 ]);
 
 export function loadFallbackConfig(): FallbackConfig {
diff --git a/src/providers/embedding/index.ts b/src/providers/embedding/index.ts
index 4580232..06417e6 100644
--- a/src/providers/embedding/index.ts
+++ b/src/providers/embedding/index.ts
@@ -1,10 +1,20 @@
 import type { EmbeddingProvider } from "../../types.js";
 import { detectEmbeddingProvider, getEnvVar } from "../../config.js";
+
+function requireEnvVar(key: string): string {
+  const value = getEnvVar(key);
+  if (!value) {
+    throw new Error(
+      `Missing required environment variable: ${key}. Set it in ~/.agentmemory/.env or as an environment variable.`,
+    );
+  }
+  return value;
+}
+
 import { GeminiEmbeddingProvider } from "./gemini.js";
 import { OpenAIEmbeddingProvider } from "./openai.js";
 import { VoyageEmbeddingProvider } from "./voyage.js";
 import { CohereEmbeddingProvider } from "./cohere.js";
-import { OpenRouterEmbeddingProvider } from "./openrouter.js";
 import { LocalEmbeddingProvider } from "./local.js";
 import { ClipEmbeddingProvider } from "./clip.js";
 
@@ -13,7 +23,6 @@ export {
   OpenAIEmbeddingProvider,
   VoyageEmbeddingProvider,
   CohereEmbeddingProvider,
-  OpenRouterEmbeddingProvider,
   LocalEmbeddingProvider,
   ClipEmbeddingProvider,
 };
@@ -35,13 +44,27 @@ export function createEmbeddingProvider(): EmbeddingProvider | null {
     case "gemini":
       return new GeminiEmbeddingProvider(getEnvVar("GEMINI_API_KEY")!);
     case "openai":
-      return new OpenAIEmbeddingProvider(getEnvVar("OPENAI_API_KEY")!);
+      return new OpenAIEmbeddingProvider(
+        getEnvVar("OPENAI_API_KEY"),
+        getEnvVar("OPENAI_EMBEDDING_BASE_URL"),
+        getEnvVar("OPENAI_EMBEDDING_MODEL"),
+      );
+    case "ollama":
+      return new OpenAIEmbeddingProvider(
+        "no-key-required",
+        getEnvVar("OLLAMA_EMBEDDING_BASE_URL") || "http://localhost:11434/v1/embeddings",
+        getEnvVar("OLLAMA_EMBEDDING_MODEL") || "llama3",
+      );
     case "voyage":
       return new VoyageEmbeddingProvider(getEnvVar("VOYAGE_API_KEY")!);
     case "cohere":
       return new CohereEmbeddingProvider(getEnvVar("COHERE_API_KEY")!);
     case "openrouter":
-      return new OpenRouterEmbeddingProvider(getEnvVar("OPENROUTER_API_KEY")!);
+      return new OpenAIEmbeddingProvider(
+        requireEnvVar("OPENROUTER_API_KEY"),
+        "https://openrouter.ai/api/v1/embeddings",
+        getEnvVar("OPENROUTER_EMBEDDING_MODEL") || "openai/text-embedding-3-small",
+      );
     case "local":
       return new LocalEmbeddingProvider();
     default:
diff --git a/src/providers/embedding/openai.ts b/src/providers/embedding/openai.ts
index c868d4a..7a5bcb8 100644
--- a/src/providers/embedding/openai.ts
+++ b/src/providers/embedding/openai.ts
@@ -1,16 +1,17 @@
 import type { EmbeddingProvider } from "../../types.js";
 import { getEnvVar } from "../../config.js";
 
-const API_URL = "https://api.openai.com/v1/embeddings";
-
 export class OpenAIEmbeddingProvider implements EmbeddingProvider {
   readonly name = "openai";
   readonly dimensions = 1536;
   private apiKey: string;
+  private baseUrl: string;
+  private model: string;
 
-  constructor(apiKey?: string) {
-    this.apiKey = apiKey || getEnvVar("OPENAI_API_KEY") || "";
-    if (!this.apiKey) throw new Error("OPENAI_API_KEY is required");
+  constructor(apiKey?: string, baseUrl?: string, model?: string) {
+    this.apiKey = apiKey || getEnvVar("OPENAI_API_KEY") || "no-key-required";
+    this.baseUrl = baseUrl || "https://api.openai.com/v1/embeddings";
+    this.model = model || "text-embedding-3-small";
   }
 
   async embed(text: string): Promise<Float32Array> {
@@ -19,14 +20,16 @@ export class OpenAIEmbeddingProvider implements EmbeddingProvider {
   }
 
   async embedBatch(texts: string[]): Promise<Float32Array[]> {
-    const response = await fetch(API_URL, {
+    const response = await fetch(this.baseUrl, {
       method: "POST",
       headers: {
-        Authorization: `Bearer ${this.apiKey}`,
+        ...(this.apiKey && this.apiKey !== "no-key-required"
+          ? { Authorization: `Bearer ${this.apiKey}` }
+          : {}),
         "Content-Type": "application/json",
       },
       body: JSON.stringify({
-        model: "text-embedding-3-small",
+        model: this.model,
         input: texts,
       }),
     });
diff --git a/src/providers/embedding/openrouter.ts b/src/providers/embedding/openrouter.ts
deleted file mode 100644
index 7d4acb2..0000000
--- a/src/providers/embedding/openrouter.ts
+++ /dev/null
@@ -1,51 +0,0 @@
-import type { EmbeddingProvider } from "../../types.js";
-import { getEnvVar } from "../../config.js";
-
-const API_URL = "https://openrouter.ai/api/v1/embeddings";
-
-export class OpenRouterEmbeddingProvider implements EmbeddingProvider {
-  readonly name = "openrouter";
-  readonly dimensions = 1536;
-  private apiKey: string;
-  private model: string;
-
-  constructor(apiKey?: string) {
-    this.apiKey = apiKey || getEnvVar("OPENROUTER_API_KEY") || "";
-    if (!this.apiKey) throw new Error("OPENROUTER_API_KEY is required");
-    this.model =
-      getEnvVar("OPENROUTER_EMBEDDING_MODEL") ||
-      "openai/text-embedding-3-small";
-  }
-
-  async embed(text: string): Promise<Float32Array> {
-    const [result] = await this.embedBatch([text]);
-    return result;
-  }
-
-  async embedBatch(texts: string[]): Promise<Float32Array[]> {
-    const response = await fetch(API_URL, {
-      method: "POST",
-      headers: {
-        Authorization: `Bearer ${this.apiKey}`,
-        "Content-Type": "application/json",
-      },
-      body: JSON.stringify({
-        model: this.model,
-        input: texts,
-      }),
-    });
-
-    if (!response.ok) {
-      const err = await response.text();
-      throw new Error(
-        `OpenRouter embedding failed (${response.status}): ${err}`,
-      );
-    }
-
-    const data = (await response.json()) as {
-      data: Array<{ embedding: number[] }>;
-    };
-
-    return data.data.map((d) => new Float32Array(d.embedding));
-  }
-}
diff --git a/src/providers/index.ts b/src/providers/index.ts
index c9e09b4..7cabda4 100644
--- a/src/providers/index.ts
+++ b/src/providers/index.ts
@@ -6,7 +6,7 @@ import type {
 import { AgentSDKProvider } from "./agent-sdk.js";
 import { AnthropicProvider } from "./anthropic.js";
 import { MinimaxProvider } from "./minimax.js";
-import { OpenRouterProvider } from "./openrouter.js";
+import { OpenAIProvider } from "./openai.js";
 import { ResilientProvider } from "./resilient.js";
 import { FallbackChainProvider } from "./fallback-chain.js";
 import { getEnvVar } from "../config.js";
@@ -58,6 +58,47 @@ export function createFallbackProvider(
 
 function createBaseProvider(config: ProviderConfig): MemoryProvider {
   switch (config.provider) {
+    case "openai":
+      return new OpenAIProvider(
+        "openai",
+        getEnvVar("OPENAI_API_KEY") || "no-key-required",
+        config.model,
+        config.maxTokens,
+        config.baseURL || "https://api.openai.com/v1/chat/completions",
+      );
+    case "ollama":
+      return new OpenAIProvider(
+        "ollama",
+        "no-key-required",
+        config.model,
+        config.maxTokens,
+        config.baseURL || "http://localhost:11434/v1/chat/completions",
+      );
+    case "vllm":
+      return new OpenAIProvider(
+        "vllm",
+        "no-key-required",
+        config.model,
+        config.maxTokens,
+        config.baseURL!,
+      );
+    case "lmstudio":
+      return new OpenAIProvider(
+        "lmstudio",
+        "no-key-required",
+        config.model,
+        config.maxTokens,
+        config.baseURL || "http://localhost:1234/v1/chat/completions",
+      );
+    case "openrouter":
+      return new OpenAIProvider(
+        "openrouter",
+        requireEnvVar("OPENROUTER_API_KEY"),
+        config.model,
+        config.maxTokens,
+        "https://openrouter.ai/api/v1/chat/completions",
+        { "HTTP-Referer": "https://github.com/rohitg00/agentmemory" },
+      );
     case "minimax":
       return new MinimaxProvider(
         requireEnvVar("MINIMAX_API_KEY"),
@@ -79,20 +120,14 @@ function createBaseProvider(config: ProviderConfig): MemoryProvider {
           "GEMINI_API_KEY (or GOOGLE_API_KEY) is required for the gemini provider",
         );
       }
-      return new OpenRouterProvider(
+      return new OpenAIProvider(
+        "gemini",
         geminiKey,
         config.model,
         config.maxTokens,
         "https://generativelanguage.googleapis.com/v1beta/openai/chat/completions",
       );
     }
-    case "openrouter":
-      return new OpenRouterProvider(
-        requireEnvVar("OPENROUTER_API_KEY"),
-        config.model,
-        config.maxTokens,
-        "https://openrouter.ai/api/v1/chat/completions",
-      );
     case "agent-sdk":
     default:
       return new AgentSDKProvider();
diff --git a/src/providers/openrouter.ts b/src/providers/openai.ts
similarity index 57%
rename from src/providers/openrouter.ts
rename to src/providers/openai.ts
index 219ce42..1c48268 100644
--- a/src/providers/openrouter.ts
+++ b/src/providers/openai.ts
@@ -1,24 +1,18 @@
 import type { MemoryProvider } from "../types.js";
 
-export class OpenRouterProvider implements MemoryProvider {
-  name: string;
-  private apiKey: string;
-  private model: string;
-  private maxTokens: number;
-  private baseUrl: string;
-
+/**
+ * Generic OpenAI-compatible provider.
+ * Works with OpenAI, LM Studio, Ollama, vLLM, Groq, OpenRouter, etc.
+ */
+export class OpenAIProvider implements MemoryProvider {
   constructor(
-    apiKey: string,
-    model: string,
-    maxTokens: number,
-    baseUrl: string,
-  ) {
-    this.apiKey = apiKey;
-    this.model = model;
-    this.maxTokens = maxTokens;
-    this.baseUrl = baseUrl;
-    this.name = baseUrl.includes("openrouter") ? "openrouter" : "gemini";
-  }
+    public name: string,
+    private apiKey: string,
+    private model: string,
+    private maxTokens: number,
+    private baseUrl: string,
+    private extraHeaders: Record<string, string> = {},
+  ) {}
 
   async compress(systemPrompt: string, userPrompt: string): Promise<string> {
     return this.call(systemPrompt, userPrompt);
@@ -36,10 +30,10 @@ export class OpenRouterProvider implements MemoryProvider {
       method: "POST",
       headers: {
         "Content-Type": "application/json",
-        Authorization: `Bearer ${this.apiKey}`,
-        ...(this.baseUrl.includes("openrouter")
-          ? { "HTTP-Referer": "https://github.com/rohitg00/agentmemory" }
+        ...(this.apiKey && this.apiKey !== "no-key-required"
+          ? { Authorization: `Bearer ${this.apiKey}` }
           : {}),
+        ...this.extraHeaders,
       },
       body: JSON.stringify({
         model: this.model,
@@ -56,11 +50,10 @@ export class OpenRouterProvider implements MemoryProvider {
       throw new Error(`${this.name} API error (${response.status}): ${text}`);
     }
 
-    const data = (await response.json()) as Record<string, unknown>;
-    const choices = data.choices as
-      | Array<{ message: { content: string } }>
-      | undefined;
-    const content = choices?.[0]?.message?.content;
+    const data = (await response.json()) as {
+      choices?: Array<{ message?: { content?: string } }>;
+    };
+    const content = data.choices?.[0]?.message?.content;
     if (!content) {
       throw new Error(
         `${this.name} returned unexpected response: ${JSON.stringify(data).slice(0, 200)}`,
diff --git a/src/types.ts b/src/types.ts
index 5a450b9..97829c8 100644
--- a/src/types.ts
+++ b/src/types.ts
@@ -127,7 +127,7 @@ export interface ProviderConfig {
   baseURL?: string;
 }
 
-export type ProviderType = "agent-sdk" | "anthropic" | "gemini" | "openrouter" | "minimax";
+export type ProviderType = "agent-sdk" | "anthropic" | "gemini" | "openrouter" | "minimax" | "lmstudio" | "openai" | "ollama" | "vllm";
 
 export interface MemoryProvider {
   name: string;
diff --git a/test/embedding-provider.test.ts b/test/embedding-provider.test.ts
index 05d01bd..139b0b0 100644
--- a/test/embedding-provider.test.ts
+++ b/test/embedding-provider.test.ts
@@ -2,47 +2,50 @@ import { describe, it, expect, vi, beforeEach, afterEach } from "vitest";
 import { createEmbeddingProvider } from "../src/providers/embedding/index.js";
 import { GeminiEmbeddingProvider } from "../src/providers/embedding/gemini.js";
 import { OpenAIEmbeddingProvider } from "../src/providers/embedding/openai.js";
+import * as config from "../src/config.js";
+
+vi.mock("../src/config.js", async (importOriginal) => {
+  const actual = await importOriginal<any>();
+  return {
+    ...actual,
+    detectEmbeddingProvider: vi.fn(),
+    getEnvVar: vi.fn(),
+  };
+});
 
 describe("createEmbeddingProvider", () => {
-  const originalEnv = { ...process.env };
+  const mockDetect = config.detectEmbeddingProvider as any;
+  const mockGetEnvVar = config.getEnvVar as any;
 
   beforeEach(() => {
-    process.env = { ...originalEnv };
-    delete process.env["GEMINI_API_KEY"];
-    delete process.env["OPENAI_API_KEY"];
-    delete process.env["VOYAGE_API_KEY"];
-    delete process.env["COHERE_API_KEY"];
-    delete process.env["OPENROUTER_API_KEY"];
-    delete process.env["EMBEDDING_PROVIDER"];
-  });
-
-  afterEach(() => {
-    process.env = originalEnv;
+    vi.resetAllMocks();
   });
 
-  it("returns null when no API keys are set", () => {
+  it("returns null when no provider is detected", () => {
+    mockDetect.mockReturnValue(null);
     const provider = createEmbeddingProvider();
     expect(provider).toBeNull();
   });
 
-  it("returns GeminiEmbeddingProvider when GEMINI_API_KEY is set", () => {
-    process.env["GEMINI_API_KEY"] = "test-key-123";
+  it("returns GeminiEmbeddingProvider when gemini is detected", () => {
+    mockDetect.mockReturnValue("gemini");
+    mockGetEnvVar.mockReturnValue("test-key");
     const provider = createEmbeddingProvider();
     expect(provider).toBeInstanceOf(GeminiEmbeddingProvider);
     expect(provider!.name).toBe("gemini");
   });
 
-  it("returns OpenAIEmbeddingProvider when OPENAI_API_KEY is set", () => {
-    process.env["OPENAI_API_KEY"] = "test-key-456";
+  it("returns OpenAIEmbeddingProvider when openai is detected", () => {
+    mockDetect.mockReturnValue("openai");
+    mockGetEnvVar.mockReturnValue("test-key");
     const provider = createEmbeddingProvider();
     expect(provider).toBeInstanceOf(OpenAIEmbeddingProvider);
     expect(provider!.name).toBe("openai");
   });
 
-  it("EMBEDDING_PROVIDER override takes precedence", () => {
-    process.env["GEMINI_API_KEY"] = "test-key-123";
-    process.env["OPENAI_API_KEY"] = "test-key-456";
-    process.env["EMBEDDING_PROVIDER"] = "openai";
+  it("uses the detected provider", () => {
+    mockDetect.mockReturnValue("openai");
+    mockGetEnvVar.mockReturnValue("test-key");
     const provider = createEmbeddingProvider();
     expect(provider).toBeInstanceOf(OpenAIEmbeddingProvider);
   });
diff --git a/test/local-providers.test.ts b/test/local-providers.test.ts
new file mode 100644
index 0000000..487d362
--- /dev/null
+++ b/test/local-providers.test.ts
@@ -0,0 +1,158 @@
+import { describe, it, expect, vi, beforeEach, afterEach } from "vitest";
+import { loadConfig, detectEmbeddingProvider } from "../src/config.js";
+import { createProvider } from "../src/providers/index.js";
+import { OpenAIProvider } from "../src/providers/openai.js";
+import { OpenAIEmbeddingProvider } from "../src/providers/embedding/openai.js";
+import { ResilientProvider } from "../src/providers/resilient.js";
+
+describe("Local Providers", () => {
+  const originalEnv = { ...process.env };
+
+  beforeEach(() => {
+    vi.resetModules();
+    process.env = { ...originalEnv };
+    delete process.env["LMSTUDIO_BASE_URL"];
+    delete process.env["LMSTUDIO_MODEL"];
+    delete process.env["OLLAMA_BASE_URL"];
+    delete process.env["OLLAMA_MODEL"];
+    delete process.env["VLLM_BASE_URL"];
+    delete process.env["VLLM_MODEL"];
+    delete process.env["OPENAI_BASE_URL"];
+    delete process.env["OPENAI_MODEL"];
+    delete process.env["ANTHROPIC_API_KEY"];
+    delete process.env["GEMINI_API_KEY"];
+    delete process.env["GOOGLE_API_KEY"];
+    delete process.env["OPENROUTER_API_KEY"];
+    delete process.env["MINIMAX_API_KEY"];
+    delete process.env["LMSTUDIO_EMBEDDING_BASE_URL"];
+    delete process.env["OLLAMA_EMBEDDING_BASE_URL"];
+    delete process.env["OLLAMA_EMBEDDING_MODEL"];
+  });
+
+  afterEach(() => {
+    process.env = originalEnv;
+  });
+
+  it("detects lmstudio provider when LMSTUDIO_BASE_URL is set", () => {
+    process.env["LMSTUDIO_BASE_URL"] = "http://localhost:1234/v1/chat/completions";
+    const config = loadConfig();
+    expect(config.provider.provider).toBe("lmstudio");
+    expect(config.provider.baseURL).toBe("http://localhost:1234/v1/chat/completions");
+  });
+
+  it("detects ollama provider when OLLAMA_MODEL is set", () => {
+    process.env["OLLAMA_MODEL"] = "llama3";
+    const config = loadConfig();
+    expect(config.provider.provider).toBe("ollama");
+    expect(config.provider.model).toBe("llama3");
+    expect(config.provider.baseURL).toBe("http://localhost:11434/v1/chat/completions");
+  });
+
+  it("detects vllm provider when VLLM_BASE_URL is set", () => {
+    process.env["VLLM_BASE_URL"] = "http://vllm-server:8000/v1/chat/completions";
+    const config = loadConfig();
+    expect(config.provider.provider).toBe("vllm");
+    expect(config.provider.baseURL).toBe("http://vllm-server:8000/v1/chat/completions");
+  });
+
+  it("detects lmstudio embedding provider when LMSTUDIO_EMBEDDING_BASE_URL is set", () => {
+    process.env["LMSTUDIO_EMBEDDING_BASE_URL"] = "http://localhost:1234/v1/embeddings";
+    const provider = detectEmbeddingProvider(process.env);
+    expect(provider).toBe("lmstudio");
+  });
+
+  it("detects ollama embedding provider when OLLAMA_EMBEDDING_MODEL is set", () => {
+    process.env["OLLAMA_EMBEDDING_MODEL"] = "nomic-embed-text";
+    const provider = detectEmbeddingProvider(process.env);
+    expect(provider).toBe("ollama");
+  });
+
+  it("creates OpenAIProvider from config for lmstudio", () => {
+    const provider = createProvider({
+      provider: "lmstudio",
+      model: "local-model",
+      maxTokens: 1000,
+      baseURL: "http://localhost:1234/v1/chat/completions",
+    });
+    expect(provider).toBeInstanceOf(ResilientProvider);
+  });
+});
+
+describe("OpenAIProvider implementation", () => {
+  it("calls fetch with correct parameters", async () => {
+    const mockFetch = vi.fn().mockResolvedValue({
+      ok: true,
+      json: async () => ({
+        choices: [{ message: { content: "mock response" } }],
+      }),
+    });
+    global.fetch = mockFetch;
+
+    const provider = new OpenAIProvider(
+      "test-provider",
+      "test-key",
+      "test-model",
+      100,
+      "http://test-url/v1/chat/completions"
+    );
+    const result = await provider.compress("system", "user");
+
+    expect(result).toBe("mock response");
+    expect(mockFetch).toHaveBeenCalledWith(
+      "http://test-url/v1/chat/completions",
+      expect.objectContaining({
+        method: "POST",
+        headers: {
+          "Content-Type": "application/json",
+          "Authorization": "Bearer test-key",
+        },
+        body: JSON.stringify({
+          model: "test-model",
+          max_tokens: 100,
+          messages: [
+            { role: "system", content: "system" },
+            { role: "user", content: "user" },
+          ],
+        }),
+      })
+    );
+  });
+});
+
+describe("OpenAIEmbeddingProvider implementation (Local compatibility)", () => {
+  it("calls fetch with correct parameters for local models", async () => {
+    const mockFetch = vi.fn().mockResolvedValue({
+      ok: true,
+      json: async () => ({
+        data: [{ embedding: [0.1, 0.2, 0.3] }],
+      }),
+    });
+    global.fetch = mockFetch;
+
+    const provider = new OpenAIEmbeddingProvider(
+      "no-key-required",
+      "http://local-ollama:11434/v1/embeddings",
+      "nomic-embed-text"
+    );
+    const result = await provider.embed("test text");
+
+    expect(result).toBeInstanceOf(Float32Array);
+    const resultArr = Array.from(result);
+    expect(resultArr[0]).toBeCloseTo(0.1);
+    expect(resultArr[1]).toBeCloseTo(0.2);
+    expect(resultArr[2]).toBeCloseTo(0.3);
+    expect(mockFetch).toHaveBeenCalledWith(
+      "http://local-ollama:11434/v1/embeddings",
+      expect.objectContaining({
+        method: "POST",
+        headers: {
+          "Content-Type": "application/json",
+        },
+        body: JSON.stringify({
+          model: "nomic-embed-text",
+          input: ["test text"],
+        }),
+      })
+    );
+  });
+});

From 79b408ee2a5e871623e73b71644dfef211fc2847 Mon Sep 17 00:00:00 2001
From: saem <saemghani+github@gmail.com>
Date: Tue, 21 Apr 2026 18:46:47 -0700
Subject: [PATCH 02/28] remove the use of mocks

also consolidated embedding provider tests
---
 test/embedding-provider.test.ts | 63 ++++++++++++++++++++-------------
 test/local-providers.test.ts    | 17 +--------
 2 files changed, 40 insertions(+), 40 deletions(-)

diff --git a/test/embedding-provider.test.ts b/test/embedding-provider.test.ts
index 139b0b0..3df4f70 100644
--- a/test/embedding-provider.test.ts
+++ b/test/embedding-provider.test.ts
@@ -1,51 +1,66 @@
-import { describe, it, expect, vi, beforeEach, afterEach } from "vitest";
+import { describe, it, expect, beforeEach, afterEach } from "vitest";
 import { createEmbeddingProvider } from "../src/providers/embedding/index.js";
 import { GeminiEmbeddingProvider } from "../src/providers/embedding/gemini.js";
 import { OpenAIEmbeddingProvider } from "../src/providers/embedding/openai.js";
 import * as config from "../src/config.js";
 
-vi.mock("../src/config.js", async (importOriginal) => {
-  const actual = await importOriginal<any>();
-  return {
-    ...actual,
-    detectEmbeddingProvider: vi.fn(),
-    getEnvVar: vi.fn(),
-  };
-});
-
 describe("createEmbeddingProvider", () => {
-  const mockDetect = config.detectEmbeddingProvider as any;
-  const mockGetEnvVar = config.getEnvVar as any;
+  const originalEnv = { ...process.env };
 
   beforeEach(() => {
-    vi.resetAllMocks();
+    process.env = { ...originalEnv };
+    // Set to empty string to override any local .env values for testing "not set" state
+    process.env["GEMINI_API_KEY"] = "";
+    process.env["OPENAI_API_KEY"] = "";
+    process.env["VOYAGE_API_KEY"] = "";
+    process.env["COHERE_API_KEY"] = "";
+    process.env["OPENROUTER_API_KEY"] = "";
+    process.env["EMBEDDING_PROVIDER"] = "";
+    process.env["OLLAMA_EMBEDDING_BASE_URL"] = "";
+    process.env["OLLAMA_EMBEDDING_MODEL"] = "";
+    process.env["LMSTUDIO_EMBEDDING_BASE_URL"] = "";
+    process.env["LMSTUDIO_EMBEDDING_MODEL"] = "";
+  });
+
+  afterEach(() => {
+    process.env = originalEnv;
   });
 
-  it("returns null when no provider is detected", () => {
-    mockDetect.mockReturnValue(null);
+  it("returns null when no API keys are set", () => {
     const provider = createEmbeddingProvider();
     expect(provider).toBeNull();
   });
 
-  it("returns GeminiEmbeddingProvider when gemini is detected", () => {
-    mockDetect.mockReturnValue("gemini");
-    mockGetEnvVar.mockReturnValue("test-key");
+  it("returns GeminiEmbeddingProvider when GEMINI_API_KEY is set", () => {
+    process.env["GEMINI_API_KEY"] = "test-key-123";
     const provider = createEmbeddingProvider();
     expect(provider).toBeInstanceOf(GeminiEmbeddingProvider);
     expect(provider!.name).toBe("gemini");
   });
 
-  it("returns OpenAIEmbeddingProvider when openai is detected", () => {
-    mockDetect.mockReturnValue("openai");
-    mockGetEnvVar.mockReturnValue("test-key");
+  it("returns OpenAIEmbeddingProvider when OPENAI_API_KEY is set", () => {
+    process.env["OPENAI_API_KEY"] = "test-key-456";
     const provider = createEmbeddingProvider();
     expect(provider).toBeInstanceOf(OpenAIEmbeddingProvider);
     expect(provider!.name).toBe("openai");
   });
 
-  it("uses the detected provider", () => {
-    mockDetect.mockReturnValue("openai");
-    mockGetEnvVar.mockReturnValue("test-key");
+  it("detects lmstudio embedding provider when LMSTUDIO_EMBEDDING_BASE_URL is set", () => {
+    process.env["LMSTUDIO_EMBEDDING_BASE_URL"] = "http://localhost:1234/v1/embeddings";
+    const provider = config.detectEmbeddingProvider(process.env);
+    expect(provider).toBe("lmstudio");
+  });
+
+  it("detects ollama embedding provider when OLLAMA_EMBEDDING_MODEL is set", () => {
+    process.env["OLLAMA_EMBEDDING_MODEL"] = "nomic-embed-text";
+    const provider = config.detectEmbeddingProvider(process.env);
+    expect(provider).toBe("ollama");
+  });
+
+  it("EMBEDDING_PROVIDER override takes precedence", () => {
+    process.env["GEMINI_API_KEY"] = "test-key-123";
+    process.env["OPENAI_API_KEY"] = "test-key-456";
+    process.env["EMBEDDING_PROVIDER"] = "openai";
     const provider = createEmbeddingProvider();
     expect(provider).toBeInstanceOf(OpenAIEmbeddingProvider);
   });
diff --git a/test/local-providers.test.ts b/test/local-providers.test.ts
index 487d362..93da429 100644
--- a/test/local-providers.test.ts
+++ b/test/local-providers.test.ts
@@ -1,5 +1,5 @@
 import { describe, it, expect, vi, beforeEach, afterEach } from "vitest";
-import { loadConfig, detectEmbeddingProvider } from "../src/config.js";
+import { loadConfig } from "../src/config.js";
 import { createProvider } from "../src/providers/index.js";
 import { OpenAIProvider } from "../src/providers/openai.js";
 import { OpenAIEmbeddingProvider } from "../src/providers/embedding/openai.js";
@@ -24,9 +24,6 @@ describe("Local Providers", () => {
     delete process.env["GOOGLE_API_KEY"];
     delete process.env["OPENROUTER_API_KEY"];
     delete process.env["MINIMAX_API_KEY"];
-    delete process.env["LMSTUDIO_EMBEDDING_BASE_URL"];
-    delete process.env["OLLAMA_EMBEDDING_BASE_URL"];
-    delete process.env["OLLAMA_EMBEDDING_MODEL"];
   });
 
   afterEach(() => {
@@ -55,18 +52,6 @@ describe("Local Providers", () => {
     expect(config.provider.baseURL).toBe("http://vllm-server:8000/v1/chat/completions");
   });
 
-  it("detects lmstudio embedding provider when LMSTUDIO_EMBEDDING_BASE_URL is set", () => {
-    process.env["LMSTUDIO_EMBEDDING_BASE_URL"] = "http://localhost:1234/v1/embeddings";
-    const provider = detectEmbeddingProvider(process.env);
-    expect(provider).toBe("lmstudio");
-  });
-
-  it("detects ollama embedding provider when OLLAMA_EMBEDDING_MODEL is set", () => {
-    process.env["OLLAMA_EMBEDDING_MODEL"] = "nomic-embed-text";
-    const provider = detectEmbeddingProvider(process.env);
-    expect(provider).toBe("ollama");
-  });
-
   it("creates OpenAIProvider from config for lmstudio", () => {
     const provider = createProvider({
       provider: "lmstudio",

From 3627620c80bbd7f4bb9cb5dd88b6f6ad72dd4c18 Mon Sep 17 00:00:00 2001
From: saem <saemghani+github@gmail.com>
Date: Thu, 23 Apr 2026 20:20:12 -0700
Subject: [PATCH 03/28] restore openrouter

---
 src/providers/embedding/index.ts      |  6 +--
 src/providers/embedding/openrouter.ts | 45 +++++++++++++++++
 src/providers/index.ts                |  6 +--
 src/providers/openrouter.ts           | 69 +++++++++++++++++++++++++++
 4 files changed, 119 insertions(+), 7 deletions(-)
 create mode 100644 src/providers/embedding/openrouter.ts
 create mode 100644 src/providers/openrouter.ts

diff --git a/src/providers/embedding/index.ts b/src/providers/embedding/index.ts
index 06417e6..3d0dc8b 100644
--- a/src/providers/embedding/index.ts
+++ b/src/providers/embedding/index.ts
@@ -13,6 +13,7 @@ function requireEnvVar(key: string): string {
 
 import { GeminiEmbeddingProvider } from "./gemini.js";
 import { OpenAIEmbeddingProvider } from "./openai.js";
+import { OpenRouterEmbeddingProvider } from "./openrouter.js";
 import { VoyageEmbeddingProvider } from "./voyage.js";
 import { CohereEmbeddingProvider } from "./cohere.js";
 import { LocalEmbeddingProvider } from "./local.js";
@@ -60,10 +61,9 @@ export function createEmbeddingProvider(): EmbeddingProvider | null {
     case "cohere":
       return new CohereEmbeddingProvider(getEnvVar("COHERE_API_KEY")!);
     case "openrouter":
-      return new OpenAIEmbeddingProvider(
+      return new OpenRouterEmbeddingProvider(
         requireEnvVar("OPENROUTER_API_KEY"),
-        "https://openrouter.ai/api/v1/embeddings",
-        getEnvVar("OPENROUTER_EMBEDDING_MODEL") || "openai/text-embedding-3-small",
+        getEnvVar("OPENROUTER_EMBEDDING_MODEL"),
       );
     case "local":
       return new LocalEmbeddingProvider();
diff --git a/src/providers/embedding/openrouter.ts b/src/providers/embedding/openrouter.ts
new file mode 100644
index 0000000..c7e963d
--- /dev/null
+++ b/src/providers/embedding/openrouter.ts
@@ -0,0 +1,45 @@
+import type { EmbeddingProvider } from "../../types.js";
+
+export class OpenRouterEmbeddingProvider implements EmbeddingProvider {
+  readonly name = "openrouter";
+  readonly dimensions = 1536;
+  private apiKey: string;
+  private model: string;
+
+  constructor(apiKey: string, model = "openai/text-embedding-3-small") {
+    this.apiKey = apiKey;
+    this.model = model;
+  }
+
+  async embed(text: string): Promise<Float32Array> {
+    const [result] = await this.embedBatch([text]);
+    return result;
+  }
+
+  async embedBatch(texts: string[]): Promise<Float32Array[]> {
+    const response = await fetch("https://openrouter.ai/api/v1/embeddings", {
+      method: "POST",
+      headers: {
+        Authorization: `Bearer ${this.apiKey}`,
+        "Content-Type": "application/json",
+        "HTTP-Referer": "https://github.com/rohitg00/agentmemory",
+        "X-Title": "agentmemory",
+      },
+      body: JSON.stringify({
+        model: this.model,
+        input: texts,
+      }),
+    });
+
+    if (!response.ok) {
+      const err = await response.text();
+      throw new Error(`OpenRouter embedding failed (${response.status}): ${err}`);
+    }
+
+    const data = (await response.json()) as {
+      data: Array<{ embedding: number[] }>;
+    };
+
+    return data.data.map((d) => new Float32Array(d.embedding));
+  }
+}
diff --git a/src/providers/index.ts b/src/providers/index.ts
index 7cabda4..ac91246 100644
--- a/src/providers/index.ts
+++ b/src/providers/index.ts
@@ -7,6 +7,7 @@ import { AgentSDKProvider } from "./agent-sdk.js";
 import { AnthropicProvider } from "./anthropic.js";
 import { MinimaxProvider } from "./minimax.js";
 import { OpenAIProvider } from "./openai.js";
+import { OpenRouterProvider } from "./openrouter.js";
 import { ResilientProvider } from "./resilient.js";
 import { FallbackChainProvider } from "./fallback-chain.js";
 import { getEnvVar } from "../config.js";
@@ -91,13 +92,10 @@ function createBaseProvider(config: ProviderConfig): MemoryProvider {
         config.baseURL || "http://localhost:1234/v1/chat/completions",
       );
     case "openrouter":
-      return new OpenAIProvider(
-        "openrouter",
+      return new OpenRouterProvider(
         requireEnvVar("OPENROUTER_API_KEY"),
         config.model,
         config.maxTokens,
-        "https://openrouter.ai/api/v1/chat/completions",
-        { "HTTP-Referer": "https://github.com/rohitg00/agentmemory" },
       );
     case "minimax":
       return new MinimaxProvider(
diff --git a/src/providers/openrouter.ts b/src/providers/openrouter.ts
new file mode 100644
index 0000000..43025a4
--- /dev/null
+++ b/src/providers/openrouter.ts
@@ -0,0 +1,69 @@
+import type { MemoryProvider } from "../types.js";
+
+export class OpenRouterProvider implements MemoryProvider {
+  name: string;
+  private apiKey: string;
+  private model: string;
+  private maxTokens: number;
+  private baseUrl: string;
+
+  constructor(
+    apiKey: string,
+    model: string,
+    maxTokens: number,
+    baseUrl = "https://openrouter.ai/api/v1/chat/completions",
+  ) {
+    this.name = "openrouter";
+    this.apiKey = apiKey;
+    this.model = model;
+    this.maxTokens = maxTokens;
+    this.baseUrl = baseUrl;
+  }
+
+  async compress(systemPrompt: string, userPrompt: string): Promise<string> {
+    return this.call(systemPrompt, userPrompt);
+  }
+
+  async summarize(systemPrompt: string, userPrompt: string): Promise<string> {
+    return this.call(systemPrompt, userPrompt);
+  }
+
+  private async call(
+    systemPrompt: string,
+    userPrompt: string,
+  ): Promise<string> {
+    const response = await fetch(this.baseUrl, {
+      method: "POST",
+      headers: {
+        Authorization: `Bearer ${this.apiKey}`,
+        "Content-Type": "application/json",
+        "HTTP-Referer": "https://github.com/rohitg00/agentmemory",
+        "X-Title": "agentmemory",
+      },
+      body: JSON.stringify({
+        model: this.model,
+        max_tokens: this.maxTokens,
+        messages: [
+          { role: "system", content: systemPrompt },
+          { role: "user", content: userPrompt },
+        ],
+      }),
+    });
+
+    if (!response.ok) {
+      const text = await response.text();
+      throw new Error(`OpenRouter API error (${response.status}): ${text}`);
+    }
+
+    const data = (await response.json()) as {
+      choices?: Array<{ message?: { content?: string } }>;
+    };
+    const content = data.choices?.[0]?.message?.content;
+    if (!content) {
+      throw new Error(
+        `OpenRouter returned unexpected response: ${JSON.stringify(data).slice(0, 200)}`,
+      );
+    }
+    return content;
+  }
+}

From bf12968e1cf2068ec65d2e1a6e95489f766fb88c Mon Sep 17 00:00:00 2001
From: saem <saemghani+github@gmail.com>
Date: Thu, 23 Apr 2026 21:04:07 -0700
Subject: [PATCH 04/28] use actual base urls everywhere

---
 src/config.ts                    |  4 +--
 src/providers/embedding/index.ts |  8 ++++-
 src/providers/index.ts           |  6 ++--
 src/providers/openai.ts          |  3 +-
 test/embedding-provider.test.ts  |  2 +-
 test/local-providers.test.ts     | 58 ++++++++++++++++++++------------
 6 files changed, 52 insertions(+), 29 deletions(-)

diff --git a/src/config.ts b/src/config.ts
index 98d259d..faeff95 100644
--- a/src/config.ts
+++ b/src/config.ts
@@ -71,7 +71,7 @@ function detectProvider(env: Record<string, string>): ProviderConfig {
       provider: "ollama",
       model: env["OLLAMA_MODEL"] || "llama3",
       maxTokens,
-      baseURL: env["OLLAMA_BASE_URL"] || "http://localhost:11434/v1/chat/completions",
+      baseURL: env["OLLAMA_BASE_URL"] || "http://localhost:11434",
     };
   }
 
@@ -89,7 +89,7 @@ function detectProvider(env: Record<string, string>): ProviderConfig {
       provider: "openai",
       model: env["OPENAI_MODEL"] || "gpt-4o",
       maxTokens,
-      baseURL: env["OPENAI_BASE_URL"] || "https://api.openai.com/v1/chat/completions",
+      baseURL: env["OPENAI_BASE_URL"] || "https://api.openai.com",
     };
   }
 
diff --git a/src/providers/embedding/index.ts b/src/providers/embedding/index.ts
index 3d0dc8b..702fdea 100644
--- a/src/providers/embedding/index.ts
+++ b/src/providers/embedding/index.ts
@@ -53,9 +53,15 @@ export function createEmbeddingProvider(): EmbeddingProvider | null {
     case "ollama":
       return new OpenAIEmbeddingProvider(
         "no-key-required",
-        getEnvVar("OLLAMA_EMBEDDING_BASE_URL") || "http://localhost:11434/v1/embeddings",
+        getEnvVar("OLLAMA_EMBEDDING_BASE_URL") || "http://localhost:11434",
         getEnvVar("OLLAMA_EMBEDDING_MODEL") || "llama3",
       );
+    case "lmstudio":
+      return new OpenAIEmbeddingProvider(
+        "no-key-required",
+        getEnvVar("LMSTUDIO_EMBEDDING_BASE_URL") || "http://localhost:1234",
+        getEnvVar("LMSTUDIO_EMBEDDING_MODEL"),
+      );
     case "voyage":
       return new VoyageEmbeddingProvider(getEnvVar("VOYAGE_API_KEY")!);
     case "cohere":
diff --git a/src/providers/index.ts b/src/providers/index.ts
index f262e27..91b3893 100644
--- a/src/providers/index.ts
+++ b/src/providers/index.ts
@@ -66,7 +66,7 @@ function createBaseProvider(config: ProviderConfig): MemoryProvider {
         getEnvVar("OPENAI_API_KEY") || "no-key-required",
         config.model,
         config.maxTokens,
-        config.baseURL || "https://api.openai.com/v1/chat/completions",
+        config.baseURL || "https://api.openai.com",
       );
     case "ollama":
       return new OpenAIProvider(
@@ -74,7 +74,7 @@ function createBaseProvider(config: ProviderConfig): MemoryProvider {
         "no-key-required",
         config.model,
         config.maxTokens,
-        config.baseURL || "http://localhost:11434/v1/chat/completions",
+        config.baseURL || "http://localhost:11434",
       );
     case "vllm":
       return new OpenAIProvider(
@@ -90,7 +90,7 @@ function createBaseProvider(config: ProviderConfig): MemoryProvider {
         "no-key-required",
         config.model,
         config.maxTokens,
-        config.baseURL || "http://localhost:1234/v1/chat/completions",
+        config.baseURL || "http://localhost:1234",
       );
     case "openrouter":
       return new OpenRouterProvider(
diff --git a/src/providers/openai.ts b/src/providers/openai.ts
index 1c48268..4f2bbb2 100644
--- a/src/providers/openai.ts
+++ b/src/providers/openai.ts
@@ -26,7 +26,8 @@ export class OpenAIProvider implements MemoryProvider {
     systemPrompt: string,
     userPrompt: string,
   ): Promise<string> {
-    const response = await fetch(this.baseUrl, {
+    const url = `${this.baseUrl}/v1/chat/completions`;
+    const response = await fetch(url, {
       method: "POST",
       headers: {
         "Content-Type": "application/json",
diff --git a/test/embedding-provider.test.ts b/test/embedding-provider.test.ts
index a22e1c8..cb71ad3 100644
--- a/test/embedding-provider.test.ts
+++ b/test/embedding-provider.test.ts
@@ -1,4 +1,4 @@
-import { describe, it, expect, beforeEach, afterEach } from "vitest";
+import { describe, it, expect, vi, beforeEach, afterEach } from "vitest";
 import { createEmbeddingProvider } from "../src/providers/embedding/index.js";
 import { GeminiEmbeddingProvider } from "../src/providers/embedding/gemini.js";
 import { OpenAIEmbeddingProvider } from "../src/providers/embedding/openai.js";
diff --git a/test/local-providers.test.ts b/test/local-providers.test.ts
index 93da429..cafc812 100644
--- a/test/local-providers.test.ts
+++ b/test/local-providers.test.ts
@@ -1,6 +1,7 @@
 import { describe, it, expect, vi, beforeEach, afterEach } from "vitest";
 import { loadConfig } from "../src/config.js";
 import { createProvider } from "../src/providers/index.js";
+import { createEmbeddingProvider } from "../src/providers/embedding/index.js";
 import { OpenAIProvider } from "../src/providers/openai.js";
 import { OpenAIEmbeddingProvider } from "../src/providers/embedding/openai.js";
 import { ResilientProvider } from "../src/providers/resilient.js";
@@ -11,19 +12,20 @@ describe("Local Providers", () => {
   beforeEach(() => {
     vi.resetModules();
     process.env = { ...originalEnv };
-    delete process.env["LMSTUDIO_BASE_URL"];
-    delete process.env["LMSTUDIO_MODEL"];
-    delete process.env["OLLAMA_BASE_URL"];
-    delete process.env["OLLAMA_MODEL"];
-    delete process.env["VLLM_BASE_URL"];
-    delete process.env["VLLM_MODEL"];
-    delete process.env["OPENAI_BASE_URL"];
-    delete process.env["OPENAI_MODEL"];
-    delete process.env["ANTHROPIC_API_KEY"];
-    delete process.env["GEMINI_API_KEY"];
-    delete process.env["GOOGLE_API_KEY"];
-    delete process.env["OPENROUTER_API_KEY"];
-    delete process.env["MINIMAX_API_KEY"];
+    process.env["LMSTUDIO_BASE_URL"] = "";
+    process.env["LMSTUDIO_MODEL"] = "";
+    process.env["OLLAMA_BASE_URL"] = "";
+    process.env["OLLAMA_MODEL"] = "";
+    process.env["VLLM_BASE_URL"] = "";
+    process.env["VLLM_MODEL"] = "";
+    process.env["OPENAI_BASE_URL"] = "";
+    process.env["OPENAI_MODEL"] = "";
+    process.env["ANTHROPIC_API_KEY"] = "";
+    process.env["GEMINI_API_KEY"] = "";
+    process.env["GOOGLE_API_KEY"] = "";
+    process.env["OPENROUTER_API_KEY"] = "";
+    process.env["MINIMAX_API_KEY"] = "";
+    process.env["EMBEDDING_PROVIDER"] = "";
   });
 
   afterEach(() => {
@@ -31,10 +33,10 @@ describe("Local Providers", () => {
   });
 
   it("detects lmstudio provider when LMSTUDIO_BASE_URL is set", () => {
-    process.env["LMSTUDIO_BASE_URL"] = "http://localhost:1234/v1/chat/completions";
+    process.env["LMSTUDIO_BASE_URL"] = "http://localhost:1234";
     const config = loadConfig();
     expect(config.provider.provider).toBe("lmstudio");
-    expect(config.provider.baseURL).toBe("http://localhost:1234/v1/chat/completions");
+    expect(config.provider.baseURL).toBe("http://localhost:1234");
   });
 
   it("detects ollama provider when OLLAMA_MODEL is set", () => {
@@ -42,14 +44,14 @@ describe("Local Providers", () => {
     const config = loadConfig();
     expect(config.provider.provider).toBe("ollama");
     expect(config.provider.model).toBe("llama3");
-    expect(config.provider.baseURL).toBe("http://localhost:11434/v1/chat/completions");
+    expect(config.provider.baseURL).toBe("http://localhost:11434");
   });
 
   it("detects vllm provider when VLLM_BASE_URL is set", () => {
-    process.env["VLLM_BASE_URL"] = "http://vllm-server:8000/v1/chat/completions";
+    process.env["VLLM_BASE_URL"] = "http://vllm-server:8000";
     const config = loadConfig();
     expect(config.provider.provider).toBe("vllm");
-    expect(config.provider.baseURL).toBe("http://vllm-server:8000/v1/chat/completions");
+    expect(config.provider.baseURL).toBe("http://vllm-server:8000");
   });
 
   it("creates OpenAIProvider from config for lmstudio", () => {
@@ -57,10 +59,24 @@ describe("Local Providers", () => {
       provider: "lmstudio",
       model: "local-model",
       maxTokens: 1000,
-      baseURL: "http://localhost:1234/v1/chat/completions",
+      baseURL: "http://localhost:1234",
     });
     expect(provider).toBeInstanceOf(ResilientProvider);
   });
+
+  it("creates OpenAIEmbeddingProvider for lmstudio", () => {
+    process.env["LMSTUDIO_EMBEDDING_BASE_URL"] = "http://localhost:1234";
+    const provider = createEmbeddingProvider();
+    expect(provider).toBeInstanceOf(OpenAIEmbeddingProvider);
+    expect(provider!.name).toBe("openai");
+  });
+
+  it("creates OpenAIEmbeddingProvider for ollama", () => {
+    process.env["OLLAMA_EMBEDDING_BASE_URL"] = "http://localhost:11434";
+    const provider = createEmbeddingProvider();
+    expect(provider).toBeInstanceOf(OpenAIEmbeddingProvider);
+    expect(provider!.name).toBe("openai");
+  });
 });
 
 describe("OpenAIProvider implementation", () => {
@@ -78,7 +94,7 @@ describe("OpenAIProvider implementation", () => {
       "test-key",
       "test-model",
       100,
-      "http://test-url/v1/chat/completions"
+      "http://test-url"
     );
     const result = await provider.compress("system", "user");
 
@@ -116,7 +132,7 @@ describe("OpenAIEmbeddingProvider implementation (Local compatibility)", () => {
 
     const provider = new OpenAIEmbeddingProvider(
       "no-key-required",
-      "http://local-ollama:11434/v1/embeddings",
+      "http://local-ollama:11434",
       "nomic-embed-text"
     );
     const result = await provider.embed("test text");

From 5a5ac74b23c746e9a40e871ed6e2b0c59e9c32a2 Mon Sep 17 00:00:00 2001
From: saem <saemghani+github@gmail.com>
Date: Thu, 23 Apr 2026 21:04:32 -0700
Subject: [PATCH 05/28] fix fs-watcher intermitent timeout

---
 test/fs-watcher.test.ts | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/test/fs-watcher.test.ts b/test/fs-watcher.test.ts
index 21e260e..674b836 100644
--- a/test/fs-watcher.test.ts
+++ b/test/fs-watcher.test.ts
@@ -49,7 +49,7 @@ describe("FilesystemWatcher", () => {
     w.start();
     try {
       writeFileSync(join(root, "notes.md"), "hello world\n");
-      await wait(800);
+      await wait(1500);
       expect(captured.length).toBeGreaterThanOrEqual(1);
       const obs = captured[captured.length - 1];
       expect(obs.url).toBe("http://localhost:3111/agentmemory/observe");

From 909f05a5a99a8b0e198df968d630b91c855b0d5e Mon Sep 17 00:00:00 2001
From: saem <saemghani+github@gmail.com>
Date: Thu, 23 Apr 2026 21:12:07 -0700
Subject: [PATCH 06/28] restore openrouter

---
 src/providers/embedding/openrouter.ts | 20 +++++++++++++-------
 src/providers/index.ts                | 17 ++++++++---------
 src/providers/openrouter.ts           | 24 +++++++++++++-----------
 3 files changed, 34 insertions(+), 27 deletions(-)

diff --git a/src/providers/embedding/openrouter.ts b/src/providers/embedding/openrouter.ts
index c7e963d..7d4acb2 100644
--- a/src/providers/embedding/openrouter.ts
+++ b/src/providers/embedding/openrouter.ts
@@ -1,4 +1,7 @@
 import type { EmbeddingProvider } from "../../types.js";
+import { getEnvVar } from "../../config.js";
+
+const API_URL = "https://openrouter.ai/api/v1/embeddings";
 
 export class OpenRouterEmbeddingProvider implements EmbeddingProvider {
   readonly name = "openrouter";
@@ -6,9 +9,12 @@ export class OpenRouterEmbeddingProvider implements EmbeddingProvider {
   private apiKey: string;
   private model: string;
 
-  constructor(apiKey: string, model = "openai/text-embedding-3-small") {
-    this.apiKey = apiKey;
-    this.model = model;
+  constructor(apiKey?: string) {
+    this.apiKey = apiKey || getEnvVar("OPENROUTER_API_KEY") || "";
+    if (!this.apiKey) throw new Error("OPENROUTER_API_KEY is required");
+    this.model =
+      getEnvVar("OPENROUTER_EMBEDDING_MODEL") ||
+      "openai/text-embedding-3-small";
   }
 
   async embed(text: string): Promise<Float32Array> {
@@ -17,13 +23,11 @@ export class OpenRouterEmbeddingProvider implements EmbeddingProvider {
   }
 
   async embedBatch(texts: string[]): Promise<Float32Array[]> {
-    const response = await fetch("https://openrouter.ai/api/v1/embeddings", {
+    const response = await fetch(API_URL, {
       method: "POST",
       headers: {
         Authorization: `Bearer ${this.apiKey}`,
         "Content-Type": "application/json",
-        "HTTP-Referer": "https://github.com/rohitg00/agentmemory",
-        "X-Title": "agentmemory",
       },
       body: JSON.stringify({
         model: this.model,
@@ -33,7 +37,9 @@ export class OpenRouterEmbeddingProvider implements EmbeddingProvider {
 
     if (!response.ok) {
       const err = await response.text();
-      throw new Error(`OpenRouter embedding failed (${response.status}): ${err}`);
+      throw new Error(
+        `OpenRouter embedding failed (${response.status}): ${err}`,
+      );
     }
 
     const data = (await response.json()) as {
diff --git a/src/providers/index.ts b/src/providers/index.ts
index 91b3893..1602a9a 100644
--- a/src/providers/index.ts
+++ b/src/providers/index.ts
@@ -92,13 +92,6 @@ function createBaseProvider(config: ProviderConfig): MemoryProvider {
         config.maxTokens,
         config.baseURL || "http://localhost:1234",
       );
-    case "openrouter":
-      return new OpenRouterProvider(
-        requireEnvVar("OPENROUTER_API_KEY"),
-        config.model,
-        config.maxTokens,
-        "https://openrouter.ai/api/v1/chat/completions",
-      );
     case "minimax":
       return new MinimaxProvider(
         requireEnvVar("MINIMAX_API_KEY"),
@@ -120,14 +113,20 @@ function createBaseProvider(config: ProviderConfig): MemoryProvider {
           "GEMINI_API_KEY (or GOOGLE_API_KEY) is required for the gemini provider",
         );
       }
-      return new OpenAIProvider(
-        "gemini",
+      return new OpenRouterProvider(
         geminiKey,
         config.model,
         config.maxTokens,
         "https://generativelanguage.googleapis.com/v1beta/openai/chat/completions",
       );
     }
+    case "openrouter":
+      return new OpenRouterProvider(
+        requireEnvVar("OPENROUTER_API_KEY"),
+        config.model,
+        config.maxTokens,
+        "https://openrouter.ai/api/v1/chat/completions",
+      );
     case "noop":
       return new NoopProvider();
     case "agent-sdk":
diff --git a/src/providers/openrouter.ts b/src/providers/openrouter.ts
index 43025a4..219ce42 100644
--- a/src/providers/openrouter.ts
+++ b/src/providers/openrouter.ts
@@ -11,13 +11,13 @@ export class OpenRouterProvider implements MemoryProvider {
     apiKey: string,
     model: string,
     maxTokens: number,
-    baseUrl = "https://openrouter.ai/api/v1/chat/completions",
+    baseUrl: string,
   ) {
-    this.name = "openrouter";
     this.apiKey = apiKey;
     this.model = model;
     this.maxTokens = maxTokens;
     this.baseUrl = baseUrl;
+    this.name = baseUrl.includes("openrouter") ? "openrouter" : "gemini";
   }
 
   async compress(systemPrompt: string, userPrompt: string): Promise<string> {
@@ -35,10 +35,11 @@ export class OpenRouterProvider implements MemoryProvider {
     const response = await fetch(this.baseUrl, {
       method: "POST",
       headers: {
-        Authorization: `Bearer ${this.apiKey}`,
         "Content-Type": "application/json",
-        "HTTP-Referer": "https://github.com/rohitg00/agentmemory",
-        "X-Title": "agentmemory",
+        Authorization: `Bearer ${this.apiKey}`,
+        ...(this.baseUrl.includes("openrouter")
+          ? { "HTTP-Referer": "https://github.com/rohitg00/agentmemory" }
+          : {}),
       },
       body: JSON.stringify({
         model: this.model,
@@ -52,16 +53,17 @@ export class OpenRouterProvider implements MemoryProvider {
 
     if (!response.ok) {
       const text = await response.text();
-      throw new Error(`OpenRouter API error (${response.status}): ${text}`);
+      throw new Error(`${this.name} API error (${response.status}): ${text}`);
     }
 
-    const data = (await response.json()) as {
-      choices?: Array<{ message?: { content?: string } }>;
-    };
-    const content = data.choices?.[0]?.message?.content;
+    const data = (await response.json()) as Record<string, unknown>;
+    const choices = data.choices as
+      | Array<{ message: { content: string } }>
+      | undefined;
+    const content = choices?.[0]?.message?.content;
     if (!content) {
       throw new Error(
-        `OpenRouter returned unexpected response: ${JSON.stringify(data).slice(0, 200)}`,
+        `${this.name} returned unexpected response: ${JSON.stringify(data).slice(0, 200)}`,
       );
     }
     return content;

From b631a7dde4f87a80f636bb11fc9941d6a971fe7f Mon Sep 17 00:00:00 2001
From: saem <saemghani+github@gmail.com>
Date: Thu, 23 Apr 2026 21:12:19 -0700
Subject: [PATCH 07/28] remove .gemini directory

---
 .gemini/settings.json | 3 ---
 1 file changed, 3 deletions(-)
 delete mode 100644 .gemini/settings.json

diff --git a/.gemini/settings.json b/.gemini/settings.json
deleted file mode 100644
index 8715189..0000000
--- a/.gemini/settings.json
+++ /dev/null
@@ -1,3 +0,0 @@
-{
-  "general": {}
-}
\ No newline at end of file

From ebd45af9477f75d668dc599930e3c8af96938129 Mon Sep 17 00:00:00 2001
From: saem <saemghani+github@gmail.com>
Date: Thu, 23 Apr 2026 21:19:40 -0700
Subject: [PATCH 08/28] undo further openrouter changes

Also add a test because `OpenRouterEmbeddingProvider` was dropped from
the export but not detected.
---
 src/providers/embedding/index.ts | 19 +++----------------
 test/embedding-provider.test.ts  |  8 ++++++++
 2 files changed, 11 insertions(+), 16 deletions(-)

diff --git a/src/providers/embedding/index.ts b/src/providers/embedding/index.ts
index 702fdea..b20f285 100644
--- a/src/providers/embedding/index.ts
+++ b/src/providers/embedding/index.ts
@@ -1,21 +1,10 @@
 import type { EmbeddingProvider } from "../../types.js";
 import { detectEmbeddingProvider, getEnvVar } from "../../config.js";
-
-function requireEnvVar(key: string): string {
-  const value = getEnvVar(key);
-  if (!value) {
-    throw new Error(
-      `Missing required environment variable: ${key}. Set it in ~/.agentmemory/.env or as an environment variable.`,
-    );
-  }
-  return value;
-}
-
 import { GeminiEmbeddingProvider } from "./gemini.js";
 import { OpenAIEmbeddingProvider } from "./openai.js";
-import { OpenRouterEmbeddingProvider } from "./openrouter.js";
 import { VoyageEmbeddingProvider } from "./voyage.js";
 import { CohereEmbeddingProvider } from "./cohere.js";
+import { OpenRouterEmbeddingProvider } from "./openrouter.js";
 import { LocalEmbeddingProvider } from "./local.js";
 import { ClipEmbeddingProvider } from "./clip.js";
 
@@ -24,6 +13,7 @@ export {
   OpenAIEmbeddingProvider,
   VoyageEmbeddingProvider,
   CohereEmbeddingProvider,
+  OpenRouterEmbeddingProvider,
   LocalEmbeddingProvider,
   ClipEmbeddingProvider,
 };
@@ -67,10 +57,7 @@ export function createEmbeddingProvider(): EmbeddingProvider | null {
     case "cohere":
       return new CohereEmbeddingProvider(getEnvVar("COHERE_API_KEY")!);
     case "openrouter":
-      return new OpenRouterEmbeddingProvider(
-        requireEnvVar("OPENROUTER_API_KEY"),
-        getEnvVar("OPENROUTER_EMBEDDING_MODEL"),
-      );
+      return new OpenRouterEmbeddingProvider(getEnvVar("OPENROUTER_API_KEY"));
     case "local":
       return new LocalEmbeddingProvider();
     default:
diff --git a/test/embedding-provider.test.ts b/test/embedding-provider.test.ts
index cb71ad3..c9502b1 100644
--- a/test/embedding-provider.test.ts
+++ b/test/embedding-provider.test.ts
@@ -2,6 +2,7 @@ import { describe, it, expect, vi, beforeEach, afterEach } from "vitest";
 import { createEmbeddingProvider } from "../src/providers/embedding/index.js";
 import { GeminiEmbeddingProvider } from "../src/providers/embedding/gemini.js";
 import { OpenAIEmbeddingProvider } from "../src/providers/embedding/openai.js";
+import { OpenRouterEmbeddingProvider } from "../src/providers/embedding/openrouter.js";
 import * as config from "../src/config.js";
 
 describe("createEmbeddingProvider", () => {
@@ -45,6 +46,13 @@ describe("createEmbeddingProvider", () => {
     expect(provider!.name).toBe("openai");
   });
 
+  it("returns OpenRouterEmbeddingProvider when OPENROUTER_API_KEY is set", () => {
+    process.env["OPENROUTER_API_KEY"] = "test-key-789";
+    const provider = createEmbeddingProvider();
+    expect(provider).toBeInstanceOf(OpenRouterEmbeddingProvider);
+    expect(provider!.name).toBe("openrouter");
+  });
+
   it("detects lmstudio embedding provider when LMSTUDIO_EMBEDDING_BASE_URL is set", () => {
     process.env["LMSTUDIO_EMBEDDING_BASE_URL"] = "http://localhost:1234/v1/embeddings";
     const provider = config.detectEmbeddingProvider(process.env);

From 585af16bdda4b97fe52e86650fb648685773e15e Mon Sep 17 00:00:00 2001
From: saem <saemghani+github@gmail.com>
Date: Thu, 23 Apr 2026 21:20:44 -0700
Subject: [PATCH 09/28] missed a `!`

---
 src/providers/embedding/index.ts | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/providers/embedding/index.ts b/src/providers/embedding/index.ts
index b20f285..83b882f 100644
--- a/src/providers/embedding/index.ts
+++ b/src/providers/embedding/index.ts
@@ -57,7 +57,7 @@ export function createEmbeddingProvider(): EmbeddingProvider | null {
     case "cohere":
       return new CohereEmbeddingProvider(getEnvVar("COHERE_API_KEY")!);
     case "openrouter":
-      return new OpenRouterEmbeddingProvider(getEnvVar("OPENROUTER_API_KEY"));
+      return new OpenRouterEmbeddingProvider(getEnvVar("OPENROUTER_API_KEY")!);
     case "local":
       return new LocalEmbeddingProvider();
     default:

From 2e20a85f94b858c7ddc1db5e68212c446ee51a2e Mon Sep 17 00:00:00 2001
From: saem <saemghani+github@gmail.com>
Date: Thu, 23 Apr 2026 21:23:28 -0700
Subject: [PATCH 10/28] fix broken api key check

---
 src/providers/embedding/openai.ts | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/providers/embedding/openai.ts b/src/providers/embedding/openai.ts
index 760c934..7b9e5cc 100644
--- a/src/providers/embedding/openai.ts
+++ b/src/providers/embedding/openai.ts
@@ -52,7 +52,7 @@ export class OpenAIEmbeddingProvider implements EmbeddingProvider {
 
   constructor(apiKey?: string, baseUrl?: string, model?: string) {
     this.apiKey = apiKey || getEnvVar("OPENAI_API_KEY") || "";
-    if (!this.apiKey && !baseUrl) throw new Error("OPENAI_API_KEY is required");
+    if (!this.apiKey) throw new Error("OPENAI_API_KEY is required");
     this.baseUrl =
       baseUrl || getEnvVar("OPENAI_BASE_URL") || DEFAULT_BASE_URL;
     this.model =

From 13da1a1839f404eca5e4bf556a04bd234296fe76 Mon Sep 17 00:00:00 2001
From: saem <saemghani+github@gmail.com>
Date: Thu, 23 Apr 2026 21:40:51 -0700
Subject: [PATCH 11/28] missed vllm

---
 src/config.ts                    |   2 +
 src/providers/embedding/index.ts |   6 ++
 test/local-providers.test.ts     | 120 ++++++++++++++++++++-----------
 3 files changed, 85 insertions(+), 43 deletions(-)

diff --git a/src/config.ts b/src/config.ts
index faeff95..cde3854 100644
--- a/src/config.ts
+++ b/src/config.ts
@@ -214,6 +214,8 @@ export function detectEmbeddingProvider(
     return "ollama";
   if (source["LMSTUDIO_EMBEDDING_BASE_URL"] || source["LMSTUDIO_EMBEDDING_MODEL"])
     return "lmstudio";
+  if (source["VLLM_EMBEDDING_BASE_URL"] || source["VLLM_EMBEDDING_MODEL"])
+    return "vllm";
   return null;
 }
 
diff --git a/src/providers/embedding/index.ts b/src/providers/embedding/index.ts
index 83b882f..a9184cd 100644
--- a/src/providers/embedding/index.ts
+++ b/src/providers/embedding/index.ts
@@ -52,6 +52,12 @@ export function createEmbeddingProvider(): EmbeddingProvider | null {
         getEnvVar("LMSTUDIO_EMBEDDING_BASE_URL") || "http://localhost:1234",
         getEnvVar("LMSTUDIO_EMBEDDING_MODEL"),
       );
+    case "vllm":
+      return new OpenAIEmbeddingProvider(
+        "no-key-required",
+        getEnvVar("VLLM_EMBEDDING_BASE_URL"),
+        getEnvVar("VLLM_EMBEDDING_MODEL"),
+      );
     case "voyage":
       return new VoyageEmbeddingProvider(getEnvVar("VOYAGE_API_KEY")!);
     case "cohere":
diff --git a/test/local-providers.test.ts b/test/local-providers.test.ts
index cafc812..8ff27d8 100644
--- a/test/local-providers.test.ts
+++ b/test/local-providers.test.ts
@@ -1,12 +1,12 @@
 import { describe, it, expect, vi, beforeEach, afterEach } from "vitest";
-import { loadConfig } from "../src/config.js";
+import { loadConfig, detectEmbeddingProvider } from "../src/config.js";
 import { createProvider } from "../src/providers/index.js";
 import { createEmbeddingProvider } from "../src/providers/embedding/index.js";
 import { OpenAIProvider } from "../src/providers/openai.js";
 import { OpenAIEmbeddingProvider } from "../src/providers/embedding/openai.js";
 import { ResilientProvider } from "../src/providers/resilient.js";
 
-describe("Local Providers", () => {
+describe("OpenAI Family Providers", () => {
   const originalEnv = { ...process.env };
 
   beforeEach(() => {
@@ -18,6 +18,7 @@ describe("Local Providers", () => {
     process.env["OLLAMA_MODEL"] = "";
     process.env["VLLM_BASE_URL"] = "";
     process.env["VLLM_MODEL"] = "";
+    process.env["OPENAI_API_KEY"] = "";
     process.env["OPENAI_BASE_URL"] = "";
     process.env["OPENAI_MODEL"] = "";
     process.env["ANTHROPIC_API_KEY"] = "";
@@ -26,61 +27,94 @@ describe("Local Providers", () => {
     process.env["OPENROUTER_API_KEY"] = "";
     process.env["MINIMAX_API_KEY"] = "";
     process.env["EMBEDDING_PROVIDER"] = "";
+    process.env["LMSTUDIO_EMBEDDING_BASE_URL"] = "";
+    process.env["VLLM_EMBEDDING_BASE_URL"] = "";
+    process.env["OLLAMA_EMBEDDING_BASE_URL"] = "";
   });
 
   afterEach(() => {
     process.env = originalEnv;
   });
 
-  it("detects lmstudio provider when LMSTUDIO_BASE_URL is set", () => {
-    process.env["LMSTUDIO_BASE_URL"] = "http://localhost:1234";
-    const config = loadConfig();
-    expect(config.provider.provider).toBe("lmstudio");
-    expect(config.provider.baseURL).toBe("http://localhost:1234");
-  });
+  describe("Detection", () => {
+    it("detects openai provider", () => {
+      process.env["OPENAI_API_KEY"] = "sk-test";
+      const config = loadConfig();
+      expect(config.provider.provider).toBe("openai");
+    });
 
-  it("detects ollama provider when OLLAMA_MODEL is set", () => {
-    process.env["OLLAMA_MODEL"] = "llama3";
-    const config = loadConfig();
-    expect(config.provider.provider).toBe("ollama");
-    expect(config.provider.model).toBe("llama3");
-    expect(config.provider.baseURL).toBe("http://localhost:11434");
-  });
+    it("detects lmstudio provider", () => {
+      process.env["LMSTUDIO_BASE_URL"] = "http://localhost:1234";
+      const config = loadConfig();
+      expect(config.provider.provider).toBe("lmstudio");
+    });
 
-  it("detects vllm provider when VLLM_BASE_URL is set", () => {
-    process.env["VLLM_BASE_URL"] = "http://vllm-server:8000";
-    const config = loadConfig();
-    expect(config.provider.provider).toBe("vllm");
-    expect(config.provider.baseURL).toBe("http://vllm-server:8000");
-  });
+    it("detects ollama provider", () => {
+      process.env["OLLAMA_MODEL"] = "llama3";
+      const config = loadConfig();
+      expect(config.provider.provider).toBe("ollama");
+    });
+
+    it("detects vllm provider", () => {
+      process.env["VLLM_BASE_URL"] = "http://vllm-server:8000";
+      const config = loadConfig();
+      expect(config.provider.provider).toBe("vllm");
+    });
 
-  it("creates OpenAIProvider from config for lmstudio", () => {
-    const provider = createProvider({
-      provider: "lmstudio",
-      model: "local-model",
-      maxTokens: 1000,
-      baseURL: "http://localhost:1234",
+    it("detects vllm embedding provider", () => {
+      process.env["VLLM_EMBEDDING_BASE_URL"] = "http://vllm-server:8000";
+      const provider = detectEmbeddingProvider(process.env);
+      expect(provider).toBe("vllm");
     });
-    expect(provider).toBeInstanceOf(ResilientProvider);
   });
 
-  it("creates OpenAIEmbeddingProvider for lmstudio", () => {
-    process.env["LMSTUDIO_EMBEDDING_BASE_URL"] = "http://localhost:1234";
-    const provider = createEmbeddingProvider();
-    expect(provider).toBeInstanceOf(OpenAIEmbeddingProvider);
-    expect(provider!.name).toBe("openai");
+  describe("Instantiation (LLM)", () => {
+    const providers: Array<"openai" | "lmstudio" | "ollama" | "vllm"> = [
+      "openai",
+      "lmstudio",
+      "ollama",
+      "vllm",
+    ];
+
+    providers.forEach((p) => {
+      it(`creates OpenAIProvider for ${p}`, () => {
+        const provider = createProvider({
+          provider: p,
+          model: "test-model",
+          maxTokens: 1000,
+          baseURL: "http://localhost:1234",
+        });
+        expect(provider).toBeInstanceOf(ResilientProvider);
+      });
+    });
   });
 
-  it("creates OpenAIEmbeddingProvider for ollama", () => {
-    process.env["OLLAMA_EMBEDDING_BASE_URL"] = "http://localhost:11434";
-    const provider = createEmbeddingProvider();
-    expect(provider).toBeInstanceOf(OpenAIEmbeddingProvider);
-    expect(provider!.name).toBe("openai");
+  describe("Instantiation (Embedding)", () => {
+    const providers: Array<"openai" | "lmstudio" | "ollama" | "vllm"> = [
+      "openai",
+      "lmstudio",
+      "ollama",
+      "vllm",
+    ];
+
+    providers.forEach((p) => {
+      it(`creates OpenAIEmbeddingProvider for ${p}`, () => {
+        // Set necessary env vars for detection/creation
+        if (p === "openai") process.env["OPENAI_API_KEY"] = "sk-test";
+        else if (p === "lmstudio") process.env["LMSTUDIO_EMBEDDING_BASE_URL"] = "http://localhost:1234";
+        else if (p === "ollama") process.env["OLLAMA_EMBEDDING_BASE_URL"] = "http://localhost:11434";
+        else if (p === "vllm") process.env["VLLM_EMBEDDING_BASE_URL"] = "http://localhost:8000";
+
+        const provider = createEmbeddingProvider();
+        expect(provider).toBeInstanceOf(OpenAIEmbeddingProvider);
+        expect(provider!.name).toBe("openai");
+      });
+    });
   });
 });
 
 describe("OpenAIProvider implementation", () => {
-  it("calls fetch with correct parameters", async () => {
+  it("calls fetch with correct parameters and appends /v1/chat/completions", async () => {
     const mockFetch = vi.fn().mockResolvedValue({
       ok: true,
       json: async () => ({
@@ -120,8 +154,8 @@ describe("OpenAIProvider implementation", () => {
   });
 });
 
-describe("OpenAIEmbeddingProvider implementation (Local compatibility)", () => {
-  it("calls fetch with correct parameters for local models", async () => {
+describe("OpenAIEmbeddingProvider implementation", () => {
+  it("calls fetch with correct parameters and appends /v1/embeddings", async () => {
     const mockFetch = vi.fn().mockResolvedValue({
       ok: true,
       json: async () => ({
@@ -132,7 +166,7 @@ describe("OpenAIEmbeddingProvider implementation (Local compatibility)", () => {
 
     const provider = new OpenAIEmbeddingProvider(
       "no-key-required",
-      "http://local-ollama:11434",
+      "http://local-runner:11434",
       "nomic-embed-text"
     );
     const result = await provider.embed("test text");
@@ -143,7 +177,7 @@ describe("OpenAIEmbeddingProvider implementation (Local compatibility)", () => {
     expect(resultArr[1]).toBeCloseTo(0.2);
     expect(resultArr[2]).toBeCloseTo(0.3);
     expect(mockFetch).toHaveBeenCalledWith(
-      "http://local-ollama:11434/v1/embeddings",
+      "http://local-runner:11434/v1/embeddings",
       expect.objectContaining({
         method: "POST",
         headers: {

From b5c49612bf79de8fae99da02ac72038e113a0758 Mon Sep 17 00:00:00 2001
From: Saem Ghani <saemghani+github@gmail.com>
Date: Thu, 23 Apr 2026 22:08:54 -0700
Subject: [PATCH 12/28] don't try to use llama3 as an embedding model

Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
---
 src/providers/embedding/index.ts | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/providers/embedding/index.ts b/src/providers/embedding/index.ts
index a9184cd..0328272 100644
--- a/src/providers/embedding/index.ts
+++ b/src/providers/embedding/index.ts
@@ -44,7 +44,7 @@ export function createEmbeddingProvider(): EmbeddingProvider | null {
       return new OpenAIEmbeddingProvider(
         "no-key-required",
         getEnvVar("OLLAMA_EMBEDDING_BASE_URL") || "http://localhost:11434",
-        getEnvVar("OLLAMA_EMBEDDING_MODEL") || "llama3",
+        getEnvVar("OLLAMA_EMBEDDING_MODEL") || "nomic-embed-text",
       );
     case "lmstudio":
       return new OpenAIEmbeddingProvider(

From edf60f222fe895cc03909721e4db0bb69e5bede2 Mon Sep 17 00:00:00 2001
From: saem <saemghani+github@gmail.com>
Date: Thu, 23 Apr 2026 22:09:42 -0700
Subject: [PATCH 13/28] fix bad baseurl guidance in readme

---
 README.md | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/README.md b/README.md
index b33b169..a8f39d5 100644
--- a/README.md
+++ b/README.md
@@ -803,7 +803,7 @@ Create `~/.agentmemory/.env`:
 # MINIMAX_API_KEY=...
 # OPENAI_API_KEY=sk-...
 # OLLAMA_MODEL=llama3
-# LMSTUDIO_BASE_URL=http://localhost:1234/v1/chat/completions
+# LMSTUDIO_BASE_URL=http://localhost:1234
 # Opt-in Claude-subscription fallback (spawns @anthropic-ai/claude-agent-sdk);
 # leave OFF unless you understand the Stop-hook recursion risk (#149 follow-up):
 # AGENTMEMORY_ALLOW_AGENT_SDK=true
@@ -816,7 +816,7 @@ Create `~/.agentmemory/.env`:
 # OPENAI_EMBEDDING_MODEL=text-embedding-3-small
 # OPENAI_EMBEDDING_DIMENSIONS=1536        # Required when the model is not in the known-models table
 # OLLAMA_EMBEDDING_MODEL=nomic-embed-text
-# LMSTUDIO_EMBEDDING_BASE_URL=http://localhost:1234/v1/embeddings
+# LMSTUDIO_EMBEDDING_BASE_URL=http://localhost:1234
 
 # Search tuning
 # BM25_WEIGHT=0.4

From b14480f793e5a7c3584be234f9912df8495ec7e7 Mon Sep 17 00:00:00 2001
From: Saem Ghani <saemghani+github@gmail.com>
Date: Thu, 23 Apr 2026 22:14:38 -0700
Subject: [PATCH 14/28] require url and model parameters for lm studio and vllm

Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
---
 src/providers/embedding/index.ts | 23 +++++++++++++----------
 1 file changed, 13 insertions(+), 10 deletions(-)

diff --git a/src/providers/embedding/index.ts b/src/providers/embedding/index.ts
index 0328272..36302c1 100644
--- a/src/providers/embedding/index.ts
+++ b/src/providers/embedding/index.ts
@@ -47,17 +47,20 @@ export function createEmbeddingProvider(): EmbeddingProvider | null {
         getEnvVar("OLLAMA_EMBEDDING_MODEL") || "nomic-embed-text",
       );
     case "lmstudio":
-      return new OpenAIEmbeddingProvider(
-        "no-key-required",
-        getEnvVar("LMSTUDIO_EMBEDDING_BASE_URL") || "http://localhost:1234",
-        getEnvVar("LMSTUDIO_EMBEDDING_MODEL"),
-      );
+      {
+        const base = getEnvVar("LMSTUDIO_EMBEDDING_BASE_URL") || "http://localhost:1234";
+        const model = getEnvVar("LMSTUDIO_EMBEDDING_MODEL");
+        if (!model) throw new Error("LMSTUDIO_EMBEDDING_MODEL is required for the lmstudio embedding provider");
+        return new OpenAIEmbeddingProvider("no-key-required", base, model);
+      }
     case "vllm":
-      return new OpenAIEmbeddingProvider(
-        "no-key-required",
-        getEnvVar("VLLM_EMBEDDING_BASE_URL"),
-        getEnvVar("VLLM_EMBEDDING_MODEL"),
-      );
+      {
+        const base = getEnvVar("VLLM_EMBEDDING_BASE_URL");
+        const model = getEnvVar("VLLM_EMBEDDING_MODEL");
+        if (!base) throw new Error("VLLM_EMBEDDING_BASE_URL is required for the vllm embedding provider");
+        if (!model) throw new Error("VLLM_EMBEDDING_MODEL is required for the vllm embedding provider");
+        return new OpenAIEmbeddingProvider("no-key-required", base, model);
+      }
     case "voyage":
       return new VoyageEmbeddingProvider(getEnvVar("VOYAGE_API_KEY")!);
     case "cohere":

From f95cd2c05086f47d53c7eab4457852f381b337ef Mon Sep 17 00:00:00 2001
From: Saem Ghani <saemghani+github@gmail.com>
Date: Thu, 23 Apr 2026 22:19:32 -0700
Subject: [PATCH 15/28] more forgiving baseUrl handling

Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
---
 src/providers/openai.ts | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/src/providers/openai.ts b/src/providers/openai.ts
index 4f2bbb2..ebdd61e 100644
--- a/src/providers/openai.ts
+++ b/src/providers/openai.ts
@@ -26,7 +26,9 @@ export class OpenAIProvider implements MemoryProvider {
     systemPrompt: string,
     userPrompt: string,
   ): Promise<string> {
-    const url = `${this.baseUrl}/v1/chat/completions`;
+    const base = this.baseUrl.replace(/\/+$/, "");
+    const path = base.endsWith("/v1") ? "/chat/completions" : "/v1/chat/completions";
+    const url = `${base}${path}`;
     const response = await fetch(url, {
       method: "POST",
       headers: {

From 0e9338489407f01c7ac2a9dd0036e4a32800d94a Mon Sep 17 00:00:00 2001
From: saem <saemghani+github@gmail.com>
Date: Thu, 23 Apr 2026 22:23:20 -0700
Subject: [PATCH 16/28] use a timeout abort signal in OpenAIProvider

---
 src/providers/openai.ts | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/src/providers/openai.ts b/src/providers/openai.ts
index ebdd61e..64455fa 100644
--- a/src/providers/openai.ts
+++ b/src/providers/openai.ts
@@ -12,6 +12,7 @@ export class OpenAIProvider implements MemoryProvider {
     private maxTokens: number,
     private baseUrl: string,
     private extraHeaders: Record<string, string> = {},
+    private timeoutMs: number = 60_000,
   ) {}
 
   async compress(systemPrompt: string, userPrompt: string): Promise<string> {
@@ -31,6 +32,7 @@ export class OpenAIProvider implements MemoryProvider {
     const url = `${base}${path}`;
     const response = await fetch(url, {
       method: "POST",
+      signal: AbortSignal.timeout(this.timeoutMs),
       headers: {
         "Content-Type": "application/json",
         ...(this.apiKey && this.apiKey !== "no-key-required"

From e15c9b4550be11ff74d61d00c5535c3fa7585ac9 Mon Sep 17 00:00:00 2001
From: saem <saemghani+github@gmail.com>
Date: Thu, 23 Apr 2026 22:47:15 -0700
Subject: [PATCH 17/28] clean-up vllm in embedding providers test setup

---
 test/embedding-provider.test.ts | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/test/embedding-provider.test.ts b/test/embedding-provider.test.ts
index c9502b1..5616bc7 100644
--- a/test/embedding-provider.test.ts
+++ b/test/embedding-provider.test.ts
@@ -21,6 +21,8 @@ describe("createEmbeddingProvider", () => {
     process.env["OLLAMA_EMBEDDING_MODEL"] = "";
     process.env["LMSTUDIO_EMBEDDING_BASE_URL"] = "";
     process.env["LMSTUDIO_EMBEDDING_MODEL"] = "";
+    process.env["VLLM_EMBEDDING_BASE_URL"] = "";
+    process.env["VLLM_EMBEDDING_MODEL"] = "";
   });
 
   afterEach(() => {

From 22aae21e702827d7e2aa659778de90df569dbf0e Mon Sep 17 00:00:00 2001
From: saem <saemghani+github@gmail.com>
Date: Thu, 23 Apr 2026 22:47:42 -0700
Subject: [PATCH 18/28] provide all necessary params

---
 test/local-providers.test.ts | 18 ++++++++++++++----
 1 file changed, 14 insertions(+), 4 deletions(-)

diff --git a/test/local-providers.test.ts b/test/local-providers.test.ts
index 8ff27d8..d6b0b48 100644
--- a/test/local-providers.test.ts
+++ b/test/local-providers.test.ts
@@ -28,8 +28,11 @@ describe("OpenAI Family Providers", () => {
     process.env["MINIMAX_API_KEY"] = "";
     process.env["EMBEDDING_PROVIDER"] = "";
     process.env["LMSTUDIO_EMBEDDING_BASE_URL"] = "";
+    process.env["LMSTUDIO_EMBEDDING_MODEL"] = "";
     process.env["VLLM_EMBEDDING_BASE_URL"] = "";
+    process.env["VLLM_EMBEDDING_MODEL"] = "";
     process.env["OLLAMA_EMBEDDING_BASE_URL"] = "";
+    process.env["OLLAMA_EMBEDDING_MODEL"] = "";
   });
 
   afterEach(() => {
@@ -100,10 +103,17 @@ describe("OpenAI Family Providers", () => {
     providers.forEach((p) => {
       it(`creates OpenAIEmbeddingProvider for ${p}`, () => {
         // Set necessary env vars for detection/creation
-        if (p === "openai") process.env["OPENAI_API_KEY"] = "sk-test";
-        else if (p === "lmstudio") process.env["LMSTUDIO_EMBEDDING_BASE_URL"] = "http://localhost:1234";
-        else if (p === "ollama") process.env["OLLAMA_EMBEDDING_BASE_URL"] = "http://localhost:11434";
-        else if (p === "vllm") process.env["VLLM_EMBEDDING_BASE_URL"] = "http://localhost:8000";
+        if (p === "openai") {
+          process.env["OPENAI_API_KEY"] = "sk-test";
+        } else if (p === "lmstudio") {
+          process.env["LMSTUDIO_EMBEDDING_BASE_URL"] = "http://localhost:1234";
+          process.env["LMSTUDIO_EMBEDDING_MODEL"] = "test-model";
+        } else if (p === "ollama") {
+          process.env["OLLAMA_EMBEDDING_BASE_URL"] = "http://localhost:11434";
+        } else if (p === "vllm") {
+          process.env["VLLM_EMBEDDING_BASE_URL"] = "http://localhost:8000";
+          process.env["VLLM_EMBEDDING_MODEL"] = "test-model";
+        }
 
         const provider = createEmbeddingProvider();
         expect(provider).toBeInstanceOf(OpenAIEmbeddingProvider);

From 1fabbaf0fbbdac1290c806c5f578631f69ac18c8 Mon Sep 17 00:00:00 2001
From: saem <saemghani+github@gmail.com>
Date: Thu, 23 Apr 2026 22:56:58 -0700
Subject: [PATCH 19/28] openai reasoning models set correct max tokens param

---
 src/providers/openai.ts      | 24 ++++++++++++++++--------
 test/local-providers.test.ts | 23 +++++++++++++++++++++++
 2 files changed, 39 insertions(+), 8 deletions(-)

diff --git a/src/providers/openai.ts b/src/providers/openai.ts
index 64455fa..798715d 100644
--- a/src/providers/openai.ts
+++ b/src/providers/openai.ts
@@ -30,6 +30,21 @@ export class OpenAIProvider implements MemoryProvider {
     const base = this.baseUrl.replace(/\/+$/, "");
     const path = base.endsWith("/v1") ? "/chat/completions" : "/v1/chat/completions";
     const url = `${base}${path}`;
+    const isReasoningModel = this.model.startsWith("o1-") || this.model.startsWith("o3-");
+    const body: Record<string, any> = {
+      model: this.model,
+      messages: [
+        { role: "system", content: systemPrompt },
+        { role: "user", content: userPrompt },
+      ],
+    };
+
+    if (isReasoningModel) {
+      body.max_completion_tokens = this.maxTokens;
+    } else {
+      body.max_tokens = this.maxTokens;
+    }
+
     const response = await fetch(url, {
       method: "POST",
       signal: AbortSignal.timeout(this.timeoutMs),
@@ -40,14 +55,7 @@ export class OpenAIProvider implements MemoryProvider {
           : {}),
         ...this.extraHeaders,
       },
-      body: JSON.stringify({
-        model: this.model,
-        max_tokens: this.maxTokens,
-        messages: [
-          { role: "system", content: systemPrompt },
-          { role: "user", content: userPrompt },
-        ],
-      }),
+      body: JSON.stringify(body),
     });
 
     if (!response.ok) {
diff --git a/test/local-providers.test.ts b/test/local-providers.test.ts
index d6b0b48..aec9a10 100644
--- a/test/local-providers.test.ts
+++ b/test/local-providers.test.ts
@@ -162,6 +162,29 @@ describe("OpenAIProvider implementation", () => {
       })
     );
   });
+
+  it("uses max_completion_tokens for reasoning models", async () => {
+    const mockFetch = vi.fn().mockResolvedValue({
+      ok: true,
+      json: async () => ({
+        choices: [{ message: { content: "mock reasoning response" } }],
+      }),
+    });
+    global.fetch = mockFetch;
+
+    const provider = new OpenAIProvider(
+      "openai",
+      "test-key",
+      "o1-mini",
+      500,
+      "https://api.openai.com"
+    );
+    await provider.compress("system", "user");
+
+    const body = JSON.parse((mockFetch.mock.calls[0][1] as RequestInit).body as string);
+    expect(body.max_completion_tokens).toBe(500);
+    expect(body.max_tokens).toBeUndefined();
+  });
 });
 
 describe("OpenAIEmbeddingProvider implementation", () => {

From 144aa4ad6444c77f6486174c2938f3350bc37bd1 Mon Sep 17 00:00:00 2001
From: saem <saemghani+github@gmail.com>
Date: Thu, 23 Apr 2026 22:58:44 -0700
Subject: [PATCH 20/28] set default base url for vllm provider

---
 src/providers/index.ts | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/providers/index.ts b/src/providers/index.ts
index 1602a9a..9007371 100644
--- a/src/providers/index.ts
+++ b/src/providers/index.ts
@@ -82,7 +82,7 @@ function createBaseProvider(config: ProviderConfig): MemoryProvider {
         "no-key-required",
         config.model,
         config.maxTokens,
-        config.baseURL!,
+        config.baseURL || "http://localhost:8000",
       );
     case "lmstudio":
       return new OpenAIProvider(

From 7acacb90f80941d6b113dcf804755bf6d50aaf02 Mon Sep 17 00:00:00 2001
From: saem <saemghani+github@gmail.com>
Date: Thu, 23 Apr 2026 23:07:08 -0700
Subject: [PATCH 21/28] add a small wait to ensure chokidar is ready

---
 test/fs-watcher.test.ts | 1 +
 1 file changed, 1 insertion(+)

diff --git a/test/fs-watcher.test.ts b/test/fs-watcher.test.ts
index 674b836..da21d3a 100644
--- a/test/fs-watcher.test.ts
+++ b/test/fs-watcher.test.ts
@@ -48,6 +48,7 @@ describe("FilesystemWatcher", () => {
     });
     w.start();
     try {
+      await wait(100);
       writeFileSync(join(root, "notes.md"), "hello world\n");
       await wait(1500);
       expect(captured.length).toBeGreaterThanOrEqual(1);

From 6334687f5136f4f2a838ffc8b6e408ad015c3ca3 Mon Sep 17 00:00:00 2001
From: saem <saemghani+github@gmail.com>
Date: Thu, 23 Apr 2026 23:07:32 -0700
Subject: [PATCH 22/28] clean-up mocks

---
 test/local-providers.test.ts | 17 +++++++++++++----
 1 file changed, 13 insertions(+), 4 deletions(-)

diff --git a/test/local-providers.test.ts b/test/local-providers.test.ts
index aec9a10..0ba209f 100644
--- a/test/local-providers.test.ts
+++ b/test/local-providers.test.ts
@@ -124,6 +124,10 @@ describe("OpenAI Family Providers", () => {
 });
 
 describe("OpenAIProvider implementation", () => {
+  afterEach(() => {
+    vi.unstubAllGlobals();
+  });
+
   it("calls fetch with correct parameters and appends /v1/chat/completions", async () => {
     const mockFetch = vi.fn().mockResolvedValue({
       ok: true,
@@ -131,7 +135,7 @@ describe("OpenAIProvider implementation", () => {
         choices: [{ message: { content: "mock response" } }],
       }),
     });
-    global.fetch = mockFetch;
+    vi.stubGlobal("fetch", mockFetch);
 
     const provider = new OpenAIProvider(
       "test-provider",
@@ -147,17 +151,18 @@ describe("OpenAIProvider implementation", () => {
       "http://test-url/v1/chat/completions",
       expect.objectContaining({
         method: "POST",
+        signal: expect.any(AbortSignal),
         headers: {
           "Content-Type": "application/json",
           "Authorization": "Bearer test-key",
         },
         body: JSON.stringify({
           model: "test-model",
-          max_tokens: 100,
           messages: [
             { role: "system", content: "system" },
             { role: "user", content: "user" },
           ],
+          max_tokens: 100,
         }),
       })
     );
@@ -170,7 +175,7 @@ describe("OpenAIProvider implementation", () => {
         choices: [{ message: { content: "mock reasoning response" } }],
       }),
     });
-    global.fetch = mockFetch;
+    vi.stubGlobal("fetch", mockFetch);
 
     const provider = new OpenAIProvider(
       "openai",
@@ -188,6 +193,10 @@ describe("OpenAIProvider implementation", () => {
 });
 
 describe("OpenAIEmbeddingProvider implementation", () => {
+  afterEach(() => {
+    vi.unstubAllGlobals();
+  });
+
   it("calls fetch with correct parameters and appends /v1/embeddings", async () => {
     const mockFetch = vi.fn().mockResolvedValue({
       ok: true,
@@ -195,7 +204,7 @@ describe("OpenAIEmbeddingProvider implementation", () => {
         data: [{ embedding: [0.1, 0.2, 0.3] }],
       }),
     });
-    global.fetch = mockFetch;
+    vi.stubGlobal("fetch", mockFetch);
 
     const provider = new OpenAIEmbeddingProvider(
       "no-key-required",

From ac16ed5231a63921f9daaff6dd002b5841ca4dc4 Mon Sep 17 00:00:00 2001
From: saem <saemghani+github@gmail.com>
Date: Thu, 23 Apr 2026 23:12:37 -0700
Subject: [PATCH 23/28] remove unused `extraHeaders` in OpenAIProviders

---
 src/providers/openai.ts | 2 --
 1 file changed, 2 deletions(-)

diff --git a/src/providers/openai.ts b/src/providers/openai.ts
index 798715d..328b2d6 100644
--- a/src/providers/openai.ts
+++ b/src/providers/openai.ts
@@ -11,7 +11,6 @@ export class OpenAIProvider implements MemoryProvider {
     private model: string,
     private maxTokens: number,
     private baseUrl: string,
-    private extraHeaders: Record<string, string> = {},
     private timeoutMs: number = 60_000,
   ) {}
 
@@ -53,7 +52,6 @@ export class OpenAIProvider implements MemoryProvider {
         ...(this.apiKey && this.apiKey !== "no-key-required"
           ? { Authorization: `Bearer ${this.apiKey}` }
           : {}),
-        ...this.extraHeaders,
       },
       body: JSON.stringify(body),
     });

From 2238d124434c95992dcf2450c39987a14f93a1c2 Mon Sep 17 00:00:00 2001
From: saem <saemghani+github@gmail.com>
Date: Thu, 23 Apr 2026 23:17:19 -0700
Subject: [PATCH 24/28] require api key to connect to openapi proper

---
 src/config.ts                | 2 +-
 test/local-providers.test.ts | 7 +++++++
 2 files changed, 8 insertions(+), 1 deletion(-)

diff --git a/src/config.ts b/src/config.ts
index cde3854..539f228 100644
--- a/src/config.ts
+++ b/src/config.ts
@@ -84,7 +84,7 @@ function detectProvider(env: Record<string, string>): ProviderConfig {
     };
   }
 
-  if (hasRealValue(env["OPENAI_API_KEY"]) || hasRealValue(env["OPENAI_BASE_URL"])) {
+  if (hasRealValue(env["OPENAI_API_KEY"])) {
     return {
       provider: "openai",
       model: env["OPENAI_MODEL"] || "gpt-4o",
diff --git a/test/local-providers.test.ts b/test/local-providers.test.ts
index 0ba209f..7c89c39 100644
--- a/test/local-providers.test.ts
+++ b/test/local-providers.test.ts
@@ -46,6 +46,13 @@ describe("OpenAI Family Providers", () => {
       expect(config.provider.provider).toBe("openai");
     });
 
+    it("does not detect openai provider with bare base URL", () => {
+      process.env["OPENAI_BASE_URL"] = "https://my-proxy.com";
+      const config = loadConfig();
+      // Should fall back to noop or next available
+      expect(config.provider.provider).not.toBe("openai");
+    });
+
     it("detects lmstudio provider", () => {
       process.env["LMSTUDIO_BASE_URL"] = "http://localhost:1234";
       const config = loadConfig();

From 3979281872fd9bb8da7381c01cda2ea76f9d7402 Mon Sep 17 00:00:00 2001
From: saem <saemghani+github@gmail.com>
Date: Fri, 24 Apr 2026 18:54:32 -0700
Subject: [PATCH 25/28] address feedback

### **Blockers (Resolved)**
   1.  **Provider Detection Order:** I reordered the detection logic in `src/config.ts` so that legacy keys
   (`ANTHROPIC_API_KEY`, `OPENAI_API_KEY`) are checked first. This prevents existing users from being
   silently switched to local providers upon upgrading.
   2.  **Explicit Override:** I implemented the `AGENTMEMORY_PROVIDER` environment variable, which allows
   users to deliberately select a provider regardless of detection priority.
   3.  **Removed Hardcoded `"gpt-4o"`:** The default OpenAI model is now `"openai-default"`, removing the
   capability assumption and requiring users to be explicit with `OPENAI_MODEL` if they want a specific
   OpenAI model.

   ### **Cleanup (Resolved)**
   1.  **Removed Sentinel Strings:** I eliminated the magic `"no-key-required"` string. The constructors for
   `OpenAIProvider` and `OpenAIEmbeddingProvider` now accept `string | null`, and the `Authorization` header
   is only attached if a key is present.
   2.  **CHANGELOG and Rebase:** I added the required notes to the `CHANGELOG.md` under `## [Unreleased]`,
   specifically documenting the constructor changes for `OpenAIEmbeddingProvider`.

   The only item **not addressed** is the consolidation of `OpenRouter` into `OpenAIProvider`. As the
   maintainer noted this was non-blocking and could be handled in a later release, I've left both code paths
   in place to minimize risk for this merge.

   All 852 tests, including the new detection tests, are passing.
---
 CHANGELOG.md                      |  8 ++++
 src/config.ts                     | 76 ++++++++++++++++---------------
 src/providers/embedding/index.ts  |  8 ++--
 src/providers/embedding/openai.ts | 21 +++++----
 src/providers/index.ts            |  8 ++--
 src/providers/openai.ts           | 16 ++++---
 test/embedding-provider.test.ts   |  5 +-
 test/local-providers.test.ts      |  2 +-
 test/provider-detection.test.ts   | 71 +++++++++++++++++++++++++++++
 9 files changed, 151 insertions(+), 64 deletions(-)
 create mode 100644 test/provider-detection.test.ts

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 15cedf7..0bb8be9 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -6,6 +6,14 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
 
 ## [Unreleased]
 
+### Changed
+
+- `OpenAIEmbeddingProvider` constructor now accepts optional `baseUrl` + `model` args; env vars remain fallback.
+- `OpenAIProvider` and `OpenAIEmbeddingProvider` now support `null` API keys to better support local-only providers (Ollama, LM Studio, vLLM).
+- Provider detection order updated to prioritize legacy providers (`ANTHROPIC_API_KEY`, `OPENAI_API_KEY`) over new local auto-detection to prevent silent behavior flips.
+- Added `AGENTMEMORY_PROVIDER` environment variable to explicitly select a provider and skip auto-detection.
+- Default OpenAI model changed from `gpt-4o` to `openai-default` to avoid hardcoded pricing/capability assumptions.
+
 ## [0.9.2] — 2026-04-22
 
 Safety + import-pipeline patch. Kills the infinite Stop-hook recursion loop that burned Claude Pro tokens on unkeyed installs, repairs every empty viewer tab after `import-jsonl`, derives lessons and crystals automatically from imported sessions, and opens up OpenAI-compatible embedding endpoints.
diff --git a/src/config.ts b/src/config.ts
index 539f228..ab3ee52 100644
--- a/src/config.ts
+++ b/src/config.ts
@@ -48,77 +48,81 @@ function hasRealValue(v: string | undefined): v is string {
 function detectProvider(env: Record<string, string>): ProviderConfig {
   const maxTokens = parseInt(env["MAX_TOKENS"] || "4096", 10);
 
-  // MiniMax: Anthropic-compatible API, requires raw fetch to avoid SDK stainless headers
-  if (hasRealValue(env["MINIMAX_API_KEY"])) {
+  const explicitProvider = env["AGENTMEMORY_PROVIDER"];
+
+  if (explicitProvider === "openai" || (!explicitProvider && hasRealValue(env["OPENAI_API_KEY"]))) {
     return {
-      provider: "minimax",
-      model: env["MINIMAX_MODEL"] || "MiniMax-M2.7",
+      provider: "openai",
+      model: env["OPENAI_MODEL"] || "openai-default",
       maxTokens,
+      baseURL: env["OPENAI_BASE_URL"] || "https://api.openai.com",
     };
   }
 
-  if (hasRealValue(env["LMSTUDIO_BASE_URL"]) || hasRealValue(env["LMSTUDIO_MODEL"])) {
+  if (explicitProvider === "anthropic" || (!explicitProvider && hasRealValue(env["ANTHROPIC_API_KEY"]))) {
     return {
-      provider: "lmstudio",
-      model: env["LMSTUDIO_MODEL"] || "local-model",
+      provider: "anthropic",
+      model: env["ANTHROPIC_MODEL"] || "claude-sonnet-4-20250514",
       maxTokens,
-      baseURL: env["LMSTUDIO_BASE_URL"],
+      baseURL: env["ANTHROPIC_BASE_URL"],
     };
   }
 
-  if (hasRealValue(env["OLLAMA_BASE_URL"]) || hasRealValue(env["OLLAMA_MODEL"])) {
+  if (explicitProvider === "gemini" || (!explicitProvider && (hasRealValue(env["GEMINI_API_KEY"]) || hasRealValue(env["GOOGLE_API_KEY"])))) {
+    if (!hasRealValue(env["GEMINI_API_KEY"]) && hasRealValue(env["GOOGLE_API_KEY"])) {
+      process.stderr.write(
+        "[agentmemory] GOOGLE_API_KEY detected — treating as GEMINI_API_KEY. " +
+          "Set GEMINI_API_KEY in ~/.agentmemory/.env to silence this warning.\n",
+      );
+    }
     return {
-      provider: "ollama",
-      model: env["OLLAMA_MODEL"] || "llama3",
+      provider: "gemini",
+      model: env["GEMINI_MODEL"] || "gemini-2.0-flash",
       maxTokens,
-      baseURL: env["OLLAMA_BASE_URL"] || "http://localhost:11434",
     };
   }
 
-  if (hasRealValue(env["VLLM_BASE_URL"]) || hasRealValue(env["VLLM_MODEL"])) {
+  if (explicitProvider === "openrouter" || (!explicitProvider && hasRealValue(env["OPENROUTER_API_KEY"]))) {
     return {
-      provider: "vllm",
-      model: env["VLLM_MODEL"] || "local-model",
+      provider: "openrouter",
+      model: env["OPENROUTER_MODEL"] || "anthropic/claude-sonnet-4-20250514",
       maxTokens,
-      baseURL: env["VLLM_BASE_URL"],
     };
   }
 
-  if (hasRealValue(env["OPENAI_API_KEY"])) {
+  // Local/Compatible providers (moved after legacy ones to avoid flipping existing users)
+  if (explicitProvider === "minimax" || (!explicitProvider && hasRealValue(env["MINIMAX_API_KEY"]))) {
     return {
-      provider: "openai",
-      model: env["OPENAI_MODEL"] || "gpt-4o",
+      provider: "minimax",
+      model: env["MINIMAX_MODEL"] || "MiniMax-M2.7",
       maxTokens,
-      baseURL: env["OPENAI_BASE_URL"] || "https://api.openai.com",
     };
   }
 
-  if (hasRealValue(env["ANTHROPIC_API_KEY"])) {
+  if (explicitProvider === "lmstudio" || (!explicitProvider && (hasRealValue(env["LMSTUDIO_BASE_URL"]) || hasRealValue(env["LMSTUDIO_MODEL"])))) {
     return {
-      provider: "anthropic",
-      model: env["ANTHROPIC_MODEL"] || "claude-sonnet-4-20250514",
+      provider: "lmstudio",
+      model: env["LMSTUDIO_MODEL"] || "local-model",
       maxTokens,
-      baseURL: env["ANTHROPIC_BASE_URL"],
+      baseURL: env["LMSTUDIO_BASE_URL"],
     };
   }
-  if (hasRealValue(env["GEMINI_API_KEY"]) || hasRealValue(env["GOOGLE_API_KEY"])) {
-    if (!hasRealValue(env["GEMINI_API_KEY"]) && hasRealValue(env["GOOGLE_API_KEY"])) {
-      process.stderr.write(
-        "[agentmemory] GOOGLE_API_KEY detected — treating as GEMINI_API_KEY. " +
-          "Set GEMINI_API_KEY in ~/.agentmemory/.env to silence this warning.\n",
-      );
-    }
+
+  if (explicitProvider === "ollama" || (!explicitProvider && (hasRealValue(env["OLLAMA_BASE_URL"]) || hasRealValue(env["OLLAMA_MODEL"])))) {
     return {
-      provider: "gemini",
-      model: env["GEMINI_MODEL"] || "gemini-2.0-flash",
+      provider: "ollama",
+      model: env["OLLAMA_MODEL"] || "llama3",
       maxTokens,
+      baseURL: env["OLLAMA_BASE_URL"] || "http://localhost:11434",
     };
   }
-  if (hasRealValue(env["OPENROUTER_API_KEY"])) {
+
+  if (explicitProvider === "vllm" || (!explicitProvider && (hasRealValue(env["VLLM_BASE_URL"]) || hasRealValue(env["VLLM_MODEL"])))) {
     return {
-      provider: "openrouter",
-      model: env["OPENROUTER_MODEL"] || "anthropic/claude-sonnet-4-20250514",
+      provider: "vllm",
+      model: env["VLLM_MODEL"] || "local-model",
       maxTokens,
+      baseURL: env["VLLM_BASE_URL"],
     };
   }
 
diff --git a/src/providers/embedding/index.ts b/src/providers/embedding/index.ts
index 36302c1..3c93779 100644
--- a/src/providers/embedding/index.ts
+++ b/src/providers/embedding/index.ts
@@ -36,13 +36,13 @@ export function createEmbeddingProvider(): EmbeddingProvider | null {
       return new GeminiEmbeddingProvider(getEnvVar("GEMINI_API_KEY")!);
     case "openai":
       return new OpenAIEmbeddingProvider(
-        getEnvVar("OPENAI_API_KEY"),
+        getEnvVar("OPENAI_API_KEY") || null,
         getEnvVar("OPENAI_EMBEDDING_BASE_URL"),
         getEnvVar("OPENAI_EMBEDDING_MODEL"),
       );
     case "ollama":
       return new OpenAIEmbeddingProvider(
-        "no-key-required",
+        null,
         getEnvVar("OLLAMA_EMBEDDING_BASE_URL") || "http://localhost:11434",
         getEnvVar("OLLAMA_EMBEDDING_MODEL") || "nomic-embed-text",
       );
@@ -51,7 +51,7 @@ export function createEmbeddingProvider(): EmbeddingProvider | null {
         const base = getEnvVar("LMSTUDIO_EMBEDDING_BASE_URL") || "http://localhost:1234";
         const model = getEnvVar("LMSTUDIO_EMBEDDING_MODEL");
         if (!model) throw new Error("LMSTUDIO_EMBEDDING_MODEL is required for the lmstudio embedding provider");
-        return new OpenAIEmbeddingProvider("no-key-required", base, model);
+        return new OpenAIEmbeddingProvider(null, base, model);
       }
     case "vllm":
       {
@@ -59,7 +59,7 @@ export function createEmbeddingProvider(): EmbeddingProvider | null {
         const model = getEnvVar("VLLM_EMBEDDING_MODEL");
         if (!base) throw new Error("VLLM_EMBEDDING_BASE_URL is required for the vllm embedding provider");
         if (!model) throw new Error("VLLM_EMBEDDING_MODEL is required for the vllm embedding provider");
-        return new OpenAIEmbeddingProvider("no-key-required", base, model);
+        return new OpenAIEmbeddingProvider(null, base, model);
       }
     case "voyage":
       return new VoyageEmbeddingProvider(getEnvVar("VOYAGE_API_KEY")!);
diff --git a/src/providers/embedding/openai.ts b/src/providers/embedding/openai.ts
index 7b9e5cc..00552d4 100644
--- a/src/providers/embedding/openai.ts
+++ b/src/providers/embedding/openai.ts
@@ -46,13 +46,12 @@ function resolveDimensions(model: string, override: string | undefined): number
 export class OpenAIEmbeddingProvider implements EmbeddingProvider {
   readonly name = "openai";
   readonly dimensions: number;
-  private apiKey: string;
+  private apiKey: string | null;
   private baseUrl: string;
   private model: string;
 
-  constructor(apiKey?: string, baseUrl?: string, model?: string) {
-    this.apiKey = apiKey || getEnvVar("OPENAI_API_KEY") || "";
-    if (!this.apiKey) throw new Error("OPENAI_API_KEY is required");
+  constructor(apiKey?: string | null, baseUrl?: string, model?: string) {
+    this.apiKey = apiKey !== undefined ? apiKey : (getEnvVar("OPENAI_API_KEY") || null);
     this.baseUrl =
       baseUrl || getEnvVar("OPENAI_BASE_URL") || DEFAULT_BASE_URL;
     this.model =
@@ -70,14 +69,16 @@ export class OpenAIEmbeddingProvider implements EmbeddingProvider {
 
   async embedBatch(texts: string[]): Promise<Float32Array[]> {
     const url = `${this.baseUrl}/v1/embeddings`;
+    const headers: Record<string, string> = {
+      "Content-Type": "application/json",
+    };
+    if (this.apiKey) {
+      headers.Authorization = `Bearer ${this.apiKey}`;
+    }
+
     const response = await fetch(url, {
       method: "POST",
-      headers: {
-        ...(this.apiKey && this.apiKey !== "no-key-required"
-          ? { Authorization: `Bearer ${this.apiKey}` }
-          : {}),
-        "Content-Type": "application/json",
-      },
+      headers,
       body: JSON.stringify({
         model: this.model,
         input: texts,
diff --git a/src/providers/index.ts b/src/providers/index.ts
index 9007371..eac71f7 100644
--- a/src/providers/index.ts
+++ b/src/providers/index.ts
@@ -63,7 +63,7 @@ function createBaseProvider(config: ProviderConfig): MemoryProvider {
     case "openai":
       return new OpenAIProvider(
         "openai",
-        getEnvVar("OPENAI_API_KEY") || "no-key-required",
+        getEnvVar("OPENAI_API_KEY") || null,
         config.model,
         config.maxTokens,
         config.baseURL || "https://api.openai.com",
@@ -71,7 +71,7 @@ function createBaseProvider(config: ProviderConfig): MemoryProvider {
     case "ollama":
       return new OpenAIProvider(
         "ollama",
-        "no-key-required",
+        null,
         config.model,
         config.maxTokens,
         config.baseURL || "http://localhost:11434",
@@ -79,7 +79,7 @@ function createBaseProvider(config: ProviderConfig): MemoryProvider {
     case "vllm":
       return new OpenAIProvider(
         "vllm",
-        "no-key-required",
+        null,
         config.model,
         config.maxTokens,
         config.baseURL || "http://localhost:8000",
@@ -87,7 +87,7 @@ function createBaseProvider(config: ProviderConfig): MemoryProvider {
     case "lmstudio":
       return new OpenAIProvider(
         "lmstudio",
-        "no-key-required",
+        null,
         config.model,
         config.maxTokens,
         config.baseURL || "http://localhost:1234",
diff --git a/src/providers/openai.ts b/src/providers/openai.ts
index 328b2d6..df2c32a 100644
--- a/src/providers/openai.ts
+++ b/src/providers/openai.ts
@@ -7,7 +7,7 @@ import type { MemoryProvider } from "../types.js";
 export class OpenAIProvider implements MemoryProvider {
   constructor(
     public name: string,
-    private apiKey: string,
+    private apiKey: string | null,
     private model: string,
     private maxTokens: number,
     private baseUrl: string,
@@ -44,15 +44,17 @@ export class OpenAIProvider implements MemoryProvider {
       body.max_tokens = this.maxTokens;
     }
 
+    const headers: Record<string, string> = {
+      "Content-Type": "application/json",
+    };
+    if (this.apiKey) {
+      headers.Authorization = `Bearer ${this.apiKey}`;
+    }
+
     const response = await fetch(url, {
       method: "POST",
       signal: AbortSignal.timeout(this.timeoutMs),
-      headers: {
-        "Content-Type": "application/json",
-        ...(this.apiKey && this.apiKey !== "no-key-required"
-          ? { Authorization: `Bearer ${this.apiKey}` }
-          : {}),
-      },
+      headers,
       body: JSON.stringify(body),
     });
 
diff --git a/test/embedding-provider.test.ts b/test/embedding-provider.test.ts
index 5616bc7..4167b85 100644
--- a/test/embedding-provider.test.ts
+++ b/test/embedding-provider.test.ts
@@ -96,9 +96,10 @@ describe("OpenAIEmbeddingProvider", () => {
     expect(provider.dimensions).toBe(1536);
   });
 
-  it("throws when no API key is provided", () => {
+  it("does not throw when no API key is provided", () => {
     delete process.env["OPENAI_API_KEY"];
-    expect(() => new OpenAIEmbeddingProvider()).toThrow("OPENAI_API_KEY is required");
+    const provider = new OpenAIEmbeddingProvider();
+    expect(provider).toBeInstanceOf(OpenAIEmbeddingProvider);
   });
 
   it("respects OPENAI_BASE_URL env var", async () => {
diff --git a/test/local-providers.test.ts b/test/local-providers.test.ts
index 7c89c39..44587e3 100644
--- a/test/local-providers.test.ts
+++ b/test/local-providers.test.ts
@@ -214,7 +214,7 @@ describe("OpenAIEmbeddingProvider implementation", () => {
     vi.stubGlobal("fetch", mockFetch);
 
     const provider = new OpenAIEmbeddingProvider(
-      "no-key-required",
+      null,
       "http://local-runner:11434",
       "nomic-embed-text"
     );
diff --git a/test/provider-detection.test.ts b/test/provider-detection.test.ts
new file mode 100644
index 0000000..d3c838f
--- /dev/null
+++ b/test/provider-detection.test.ts
@@ -0,0 +1,71 @@
+import { describe, it, expect, vi, beforeEach, afterEach } from "vitest";
+import { loadConfig } from "../src/config.js";
+
+describe("Provider Detection Priority and Overrides", () => {
+  const originalEnv = { ...process.env };
+
+  beforeEach(() => {
+    vi.resetModules();
+    process.env = { ...originalEnv };
+    // Clear all relevant env vars
+    const vars = [
+      "OPENAI_API_KEY", "ANTHROPIC_API_KEY", "GEMINI_API_KEY", "OPENROUTER_API_KEY",
+      "MINIMAX_API_KEY", "LMSTUDIO_BASE_URL", "OLLAMA_MODEL", "VLLM_BASE_URL",
+      "AGENTMEMORY_PROVIDER", "OPENAI_MODEL"
+    ];
+    vars.forEach(v => delete process.env[v]);
+  });
+
+  afterEach(() => {
+    process.env = originalEnv;
+  });
+
+  it("prioritizes legacy providers over new local providers", () => {
+    process.env["ANTHROPIC_API_KEY"] = "sk-ant-test";
+    process.env["OLLAMA_MODEL"] = "llama3";
+    
+    const config = loadConfig();
+    expect(config.provider.provider).toBe("anthropic");
+    expect(config.provider.model).toBe("claude-sonnet-4-20250514");
+  });
+
+  it("prioritizes openai over anthropic", () => {
+    process.env["OPENAI_API_KEY"] = "sk-openai-test";
+    process.env["ANTHROPIC_API_KEY"] = "sk-ant-test";
+    
+    const config = loadConfig();
+    expect(config.provider.provider).toBe("openai");
+    expect(config.provider.model).toBe("openai-default");
+  });
+
+  it("respects AGENTMEMORY_PROVIDER override and sets correct default model", () => {
+    process.env["OPENAI_API_KEY"] = "sk-openai-test";
+    process.env["AGENTMEMORY_PROVIDER"] = "ollama";
+    
+    const config = loadConfig();
+    expect(config.provider.provider).toBe("ollama");
+    expect(config.provider.model).toBe("llama3");
+  });
+
+  it("allows AGENTMEMORY_PROVIDER without a corresponding API key", () => {
+    process.env["AGENTMEMORY_PROVIDER"] = "openai";
+    // No OPENAI_API_KEY set
+    
+    const config = loadConfig();
+    expect(config.provider.provider).toBe("openai");
+    expect(config.provider.model).toBe("openai-default");
+  });
+
+  it("uses openai-default when OPENAI_MODEL is unset", () => {
+    process.env["OPENAI_API_KEY"] = "sk-openai-test";
+    const config = loadConfig();
+    expect(config.provider.model).toBe("openai-default");
+  });
+
+  it("uses provided OPENAI_MODEL when set", () => {
+    process.env["OPENAI_API_KEY"] = "sk-openai-test";
+    process.env["OPENAI_MODEL"] = "gpt-4o-mini";
+    const config = loadConfig();
+    expect(config.provider.model).toBe("gpt-4o-mini");
+  });
+});

From ba6b298356624d74e095898972c1f57a8866f69f Mon Sep 17 00:00:00 2001
From: saem <saemghani+github@gmail.com>
Date: Fri, 24 Apr 2026 19:10:15 -0700
Subject: [PATCH 26/28] fix the changelog

---
 CHANGELOG.md | 18 +++++++++++++-----
 1 file changed, 13 insertions(+), 5 deletions(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index a437c49..e1df309 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -6,13 +6,21 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
 
 ## [Unreleased]
 
+### Added
+
+- **Support for local OpenAI-compatible providers.** Full support for Ollama, vLLM, and LM Studio as first-class providers. Added new configuration variables:
+  - **Ollama**: `OLLAMA_BASE_URL` (default: `http://localhost:11434`) and `OLLAMA_MODEL`.
+  - **LM Studio**: `LMSTUDIO_BASE_URL` and `LMSTUDIO_MODEL`.
+  - **vLLM**: `VLLM_BASE_URL` and `VLLM_MODEL`.
+  Auto-detection now supports these local endpoints while ensuring existing legacy keys (`ANTHROPIC_API_KEY`, `OPENAI_API_KEY`) remain prioritized for stable upgrades.
+- **`AGENTMEMORY_PROVIDER` environment variable.** Allows explicit selection of the LLM provider (e.g., `AGENTMEMORY_PROVIDER=ollama`), bypassing auto-detection and supporting setups with multiple keys.
+- **Improved `OpenAIEmbeddingProvider` flexibility.** The constructor now accepts optional `baseUrl` and `model` overrides, enabling direct instantiation for custom OpenAI-compatible embedding proxies.
+
 ### Changed
 
-- `OpenAIEmbeddingProvider` constructor now accepts optional `baseUrl` + `model` args; env vars remain fallback.
-- `OpenAIProvider` and `OpenAIEmbeddingProvider` now support `null` API keys to better support local-only providers (Ollama, LM Studio, vLLM).
-- Provider detection order updated to prioritize legacy providers (`ANTHROPIC_API_KEY`, `OPENAI_API_KEY`) over new local auto-detection to prevent silent behavior flips.
-- Added `AGENTMEMORY_PROVIDER` environment variable to explicitly select a provider and skip auto-detection.
-- Default OpenAI model changed from `gpt-4o` to `openai-default` to avoid hardcoded pricing/capability assumptions.
+- **Provider detection order and MiniMax priority.** Detection order updated to prioritize legacy providers (`ANTHROPIC_API_KEY`, `OPENAI_API_KEY`) over local auto-detection. `MiniMax` detection now also honors this priority and supports `MINIMAX_MODEL`.
+- **`OpenAIProvider` and `OpenAIEmbeddingProvider` key handling.** Both now support `null` API keys to natively support local-only endpoints that don't require an `Authorization` header.
+- **Neutral OpenAI default model.** The OpenAI provider now defaults to `openai-default` instead of a specific commercial model to avoid opinionated pricing or capability assumptions in the library core.
 
 ## [0.9.3] — 2026-04-24
 

From 43f6780d2f2482362068e9a030915a49df38e89f Mon Sep 17 00:00:00 2001
From: saem <saemghani+github@gmail.com>
Date: Fri, 24 Apr 2026 19:31:31 -0700
Subject: [PATCH 27/28] reasoning model detection if there are prefixes

---
 src/providers/openai.ts      |  6 +++++-
 test/local-providers.test.ts | 22 ++++++++++++++++++++++
 2 files changed, 27 insertions(+), 1 deletion(-)

diff --git a/src/providers/openai.ts b/src/providers/openai.ts
index df2c32a..e187ed9 100644
--- a/src/providers/openai.ts
+++ b/src/providers/openai.ts
@@ -29,7 +29,11 @@ export class OpenAIProvider implements MemoryProvider {
     const base = this.baseUrl.replace(/\/+$/, "");
     const path = base.endsWith("/v1") ? "/chat/completions" : "/v1/chat/completions";
     const url = `${base}${path}`;
-    const isReasoningModel = this.model.startsWith("o1-") || this.model.startsWith("o3-");
+    
+    // Detect reasoning models (o1, o3, etc) which require max_completion_tokens.
+    // We check for the pattern o1- or o3- anywhere in the string to handle prefixes.
+    const isReasoningModel = /\bo[13]-/.test(this.model);
+    
     const body: Record<string, any> = {
       model: this.model,
       messages: [
diff --git a/test/local-providers.test.ts b/test/local-providers.test.ts
index 44587e3..fbac125 100644
--- a/test/local-providers.test.ts
+++ b/test/local-providers.test.ts
@@ -197,6 +197,28 @@ describe("OpenAIProvider implementation", () => {
     expect(body.max_completion_tokens).toBe(500);
     expect(body.max_tokens).toBeUndefined();
   });
+
+  it("detects reasoning models with prefixes, in case of proxies", async () => {
+    const mockFetch = vi.fn().mockResolvedValue({
+      ok: true,
+      json: async () => ({
+        choices: [{ message: { content: "mock reasoning response" } }],
+      }),
+    });
+    vi.stubGlobal("fetch", mockFetch);
+
+    const provider = new OpenAIProvider(
+      "openai",
+      "test-key",
+      "openai/o1-mini",
+      500,
+      "https://api.openai.com"
+    );
+    await provider.compress("system", "user");
+
+    const body = JSON.parse((mockFetch.mock.calls[0][1] as RequestInit).body as string);
+    expect(body.max_completion_tokens).toBe(500);
+  });
 });
 
 describe("OpenAIEmbeddingProvider implementation", () => {

From 5994cfa781fca82956d54eee4647809f7bb03008 Mon Sep 17 00:00:00 2001
From: Saem Ghani <saemghani+github@gmail.com>
Date: Sat, 25 Apr 2026 01:41:17 -0700
Subject: [PATCH 28/28] README: note that OpenAIEmbeddingProvider env var
 fallback remain

---
 CHANGELOG.md | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index e1df309..a0f0f6d 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -14,7 +14,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
   - **vLLM**: `VLLM_BASE_URL` and `VLLM_MODEL`.
   Auto-detection now supports these local endpoints while ensuring existing legacy keys (`ANTHROPIC_API_KEY`, `OPENAI_API_KEY`) remain prioritized for stable upgrades.
 - **`AGENTMEMORY_PROVIDER` environment variable.** Allows explicit selection of the LLM provider (e.g., `AGENTMEMORY_PROVIDER=ollama`), bypassing auto-detection and supporting setups with multiple keys.
-- **Improved `OpenAIEmbeddingProvider` flexibility.** The constructor now accepts optional `baseUrl` and `model` overrides, enabling direct instantiation for custom OpenAI-compatible embedding proxies.
+- **Improved `OpenAIEmbeddingProvider` flexibility.** The constructor now accepts optional `baseUrl` and `model` overrides, enabling direct instantiation for custom OpenAI-compatible embedding proxies. Environment variables remain as fallback.
 
 ### Changed