opencode-supermemory (fork)

Fork of supermemoryai/opencode-supermemory with a fix for the Part ID schema validation crash introduced in OpenCode v1.2.25+.

What was fixed: The original plugin generates message Part IDs with supermemory-nudge-* and supermemory-context-* prefixes. OpenCode v1.2.25 (commit 090f636, PR #16966) introduced strict Zod validation requiring all Part IDs to start with prt. This caused a ZodError crash in createUserMessage() on every message. This fork generates compliant prt_-prefixed IDs.

Upstream issue: supermemoryai/opencode-supermemory#28

---

OpenCode plugin for persistent memory using Supermemory.

Your agent remembers what you tell it - across sessions, across projects.

Installation

For Humans

bunx github:pondzikk/opencode-supermemory install

Then get your API key from app.supermemory.ai and set it:

export SUPERMEMORY_API_KEY="sm_..."

Or let your agent do it - paste this into OpenCode:

Install opencode-supermemory by following https://raw.githubusercontent.com/pondzikk/opencode-supermemory/main/README.md

For LLM Agents

Instructions for LLM Agents (click to expand)

If you're an LLM agent helping set up opencode-supermemory, follow these steps:

Step 1: Run the installer

bunx github:pondzikk/opencode-supermemory install --no-tui

This will:

• Register the plugin in ~/.config/opencode/opencode.jsonc
• Create the /supermemory-init command

Step 2: Verify the config

cat ~/.config/opencode/opencode.jsonc

Should contain:

{
  "plugin": ["github:pondzikk/opencode-supermemory"]
}

If not, add it manually:

JSONC:

{
  "plugin": [
    "github:pondzikk/opencode-supermemory",
    // ... other plugins
  ],
}

JSON:

{
  "plugin": ["github:pondzikk/opencode-supermemory"]
}

Step 3: Configure API key

Ask the user to get their API key from app.supermemory.ai.

Then set it via environment variable:

export SUPERMEMORY_API_KEY="sm_..."

Or create ~/.config/opencode/supermemory.jsonc:

{
  "apiKey": "sm_...",
}

Step 4: Verify setup

Tell the user to restart OpenCode and run:

opencode -c

They should see supermemory in the tools list. If not, check:

1. Is SUPERMEMORY_API_KEY set?
2. Is the plugin in opencode.jsonc?
3. Check logs: tail ~/.opencode-supermemory.log

Step 5: Initialize codebase memory (optional)

Run /supermemory-init to have the agent explore and memorize the codebase.

Features

Context Injection

On first message, the agent receives (invisible to user):

• User profile (cross-project preferences)
• Project memories (all project knowledge)
• Relevant user memories (semantic search)

Example of what the agent sees:

[SUPERMEMORY]

User Profile:
- Prefers concise responses
- Expert in TypeScript

Project Knowledge:
- [100%] Uses Bun, not Node.js
- [100%] Build: bun run build

Relevant Memories:
- [82%] Build fails if .env.local missing

The agent uses this context automatically - no manual prompting needed.

Keyword Detection

Say "remember", "save this", "don't forget" etc. and the agent auto-saves to memory.

You: "Remember that this project uses bun"
Agent: [saves to project memory]

Add custom triggers via keywordPatterns config.

Codebase Indexing

Run /supermemory-init to explore and memorize your codebase structure, patterns, and conventions.

Preemptive Compaction

When context hits 80% capacity:

1. Triggers OpenCode's summarization
2. Injects project memories into summary context
3. Saves session summary as a memory

This preserves conversation context across compaction events.

Privacy

API key is <private>sk-abc123</private>

Content in <private> tags is never stored.

Tool Usage

The supermemory tool is available to the agent:

Mode	Args	Description
add	content, type?, scope?	Store memory
search	query, scope?	Search memories
profile	query?	View user profile
list	scope?, limit?	List memories
forget	memoryId, scope?	Delete memory

Scopes: user (cross-project), project (default)

Types: project-config, architecture, error-solution, preference, learned-pattern, conversation

Memory Scoping

Scope	Tag	Persists
User	opencode_user_{sha256(git email)}	All projects
Project	opencode_project_{sha256(directory)}	This project

Configuration

Create ~/.config/opencode/supermemory.jsonc:

{
  // API key (can also use SUPERMEMORY_API_KEY env var)
  "apiKey": "sm_...",

  // Min similarity for memory retrieval (0-1)
  "similarityThreshold": 0.6,

  // Max memories injected per request
  "maxMemories": 5,

  // Max project memories listed
  "maxProjectMemories": 10,

  // Max profile facts injected
  "maxProfileItems": 5,

  // Include user profile in context
  "injectProfile": true,

  // Prefix for container tags (used when userContainerTag/projectContainerTag not set)
  "containerTagPrefix": "opencode",

  // Optional: Set exact user container tag (overrides auto-generated tag)
  "userContainerTag": "my-custom-user-tag",

  // Optional: Set exact project container tag (overrides auto-generated tag)
  "projectContainerTag": "my-project-tag",

  // Extra keyword patterns for memory detection (regex)
  "keywordPatterns": ["log\\s+this", "write\\s+down"],

  // Context usage ratio that triggers compaction (0-1)
  "compactionThreshold": 0.8,

  // Enable semantic reranking of search results (improves relevance)
  "enableRerank": true,

  // Entity context for user memory extraction (describes the user's role/domain)
  "userEntityContext": "This is a software developer's personal context. Extract preferences, coding style, tool choices, workflow patterns, and technical decisions.",

  // Entity context for project memory extraction (describes the project domain)
  "projectEntityContext": "This is a software project's codebase context. Extract architecture decisions, file structure patterns, dependency choices, conventions, and known issues."
}

All fields optional. Env var SUPERMEMORY_API_KEY takes precedence over config file.

Container Tag Selection

By default, container tags are auto-generated using containerTagPrefix plus a hash:

• User tag: {prefix}_user_{hash(git_email)}
• Project tag: {prefix}_project_{hash(directory)}

You can override this by specifying exact container tags:

{
  // Use a specific container tag for user memories
  "userContainerTag": "my-team-workspace",

  // Use a specific container tag for project memories
  "projectContainerTag": "my-awesome-project",
}

This is useful when you want to:

• Share memories across team members (same userContainerTag)
• Sync memories between different machines for the same project
• Organize memories using your own naming scheme
• Integrate with existing Supermemory container tags from other tools

Improvements Over Upstream

1. Semantic Reranking (enableRerank: true) — Search results are reranked for better relevance using Supermemory's native rerank API
2. Stable Memory IDs (customId) — Memories get SHA-256-based custom IDs preventing duplicate entries on re-ingestion
3. Entity Context (userEntityContext, projectEntityContext) — Memories are stored with entity context descriptions that improve extraction quality
4. Native Conversation Ingestion — Multi-turn conversations use Supermemory's /v4/conversations endpoint (with fallback to structured text)
5. Soft-Delete — Memory deletion uses memories.forget() (recoverable) instead of hard delete
6. Profile Search Dedup — On context injection, if the profile API returns embedded search results, the separate searchMemories() call is skipped (saves 1 API call per message)

Note: The prt_ prefix fix is preserved from the original fork.

Usage with Oh My OpenCode

If you're using Oh My OpenCode, disable its built-in auto-compact hook to let supermemory handle context compaction:

Add to ~/.config/opencode/oh-my-opencode.json:

{
  "disabled_hooks": ["anthropic-context-window-limit-recovery"]
}

Development

bun install
bun run build
bun run typecheck

Local install:

{
  "plugin": ["file:///path/to/opencode-supermemory"],
}

Logs

tail -f ~/.opencode-supermemory.log

License

MIT