v0.1.14 — OpenRouter Raw API Migration

v0.1.14 — OpenRouter Raw API Migration

🔧 CLI / Server

• OpenRouter integration rewritten: dropped @openrouter/ai-sdk-provider dependency, migrated to raw fetch API with Snd proper token usage tracking
• Server-side OpenRouter chat endpoint now streams tool calls with tool-call typed events
• OpenRouter structured generation endpoint migrated from SDK to raw API

v0.1.12 — Model Config Tuning

v0.1.12 — Model Config Tuning

🔧 CLI / Server

• Per-model maxTokens configuration for server-side chat and object generation endpoints
• Kimi K2.6 model reference updated (removed :free suffix) — now a paid model across model lists, help docs, and context window config
• Default OpenRouter model max tokens tuned per model

v0.1.11 — Chat API Refactor & Launch Polish

v0.1.11 — Chat API Refactor & Launch Polish

🔧 CLI / Server

• Server-side chat endpoint refactored to use @openrouter/ai-sdk-provider with streamText — cleaner streaming, built-in tool loop (maxSteps: 5), and proper usage reporting
• Removed 200+ lines of manual SSE parsing and tool loop logic

🎨 Homepage

• Launch page enhanced with beta features list, coming next roadmap, and install command section
• Post-launch CTA updated to link to Quickstart guide

v0.1.10 — Beta Launch

v0.1.10 — Beta Launch

🚀 New

• Quickstart guide — Step-by-step install/login/first-prompt walkthrough
• Waitlist feature with email notifications and beta countdown
• Comparison page — Supercode vs other SWE agents
• Launch countdown and beta banner components
• /connect command — Connect AI providers from CLI
• Provider configuration — Enhanced AI model management
• Device code auth — Secure CLI login via browser

📖 Documentation

• Complete docs site redesign: typography overhaul, section-grouped sidebar, prev/next nav
• New quickstart guide
• Scroll-aware On this page TOC
• Auth and architecture pages

🎨 Homepage

• Redesigned hero with tabbed install commands
• Updated footer, navbar, OG images, changelog

🔧 CLI client

• Device approval and code entry screens
• PixelLogo component, cleaned-up login form
• Server dependency updates

Last versions updates

v0.1.8 — Waitlist & Launch Prep

• Waitlist feature with email notifications
• Beta countdown and beta banner components
• Launch page added to footer
• Open Graph images updated for social sharing
• Comparison page — Supercode vs other SWE agents with cursor animation
• Waitlist page layout cleanup

v0.1.9 — Provider & CLI Configuration

• AI model management — Enhanced model configuration and CLI setup
• Provider connections — Connect OpenAI, Anthropic, Gemini, and more
• `/connect` command — Interactive provider setup from the terminal
• CLI help documentation updates

v0.1.7

Docs

• README rewritten with new branding ("AI-Powered SWE Agent"), updated app/package table, Terminal Stack architecture section, and new badges (Next.js 16, License).
• CONTRIBUTING rewritten from scratch: Bun-first setup, both database setup steps (dashboard + terminal CLI), expanded project tree, full command reference, code style guidelines, CLI pattern doc, DB migration workflows.
• Config documentation updated with production URL guidance for OAuth setup — thanks @aviisharma238 (https://github.com/aviisharma238)!

CLI / TUI

• Phosphor CRT terminal theme — green phosphor on black, amber accent for cursor/badges, green glow highlights. New pixelWordmark() renders "SUPERCODE" as ASCII pixel art.
• Main screen redesigned with sectionHeader, cardStack, rowCard, keyValue, statusBar utilities — centered pixel wordmark, system HUD panel, command card stack, and a footer status bar.
• Chat startup overhauled: pixel wordmark header, status bar with mode/model info, /model / /help / Tab quick-start row.
• Framed "goodbye" message on exit instead of a single line.
• Animated thinking spinner with live tool call and reasoning updates (existing, refined).
• 1,048 insertions across 7 files — the TUI is now the primary surface identity.

v0.1.6

Core

• Agent mode rewritten from rigid generateApplication() to tool-calling loop (ToolLoopAgent) — model calls write_file/run_command tools directly, enabling iterative building and error recovery.
• New /model slash command — switch AI provider/model mid-session (Gemini, OpenRouter, NVIDIA NIM).
• Persistent stdin handler — prevents "terminal state may be corrupted" crashes.
• Stream cancellation via Escape key, with partial response preservation.
• Global unhandledRejection/uncaughtException handlers for crash resilience.

Tools

• Standalone tool definitions: write-file, run-command, read-instructions with path traversal protection, 1MB size limit, auto mkdir, and interactive stdin support.
• PermissionManager — granular tool permission system with wildcard matching, dangerous command detection (20 regex patterns), and [y] Once / [a] Always / [n] Deny prompts.
• ToolExecutor — manual tool-calling loop orchestrator for streaming providers.
• OpenRouter provider rewritten from SDK to raw fetch API with native tool call detection and multi-iteration tool loops.
• NVIDIA NIM provider migrated to @ai-sdk/openai-compatible SDK with proper tool-calling support.

CLI

• Model switching at runtime via /model command: Google Gemini, OpenRouter (GPT OSS, DeepSeek V4 Flash, MiniMax M3, GLM 5.1, Kimi K2.6), NVIDIA NIM (MiniMax M2.7, DeepSeek V4 Flash, Llama 3.3 70B).
• Default OpenRouter model changed to moonshotai/kimi-k2.6:free.
• Animated thinking spinner with live tool call and reasoning updates.
• Agent mode shows live tool call progress; supports exit and continue prompts.
• Chat loop resilience — catch-all error handling prevents crashes; terminal state restored on errors.

v0.1.5

Core

• New OpenRouter models: MiniMax M3, GLM 5.1, and Kimi K2.6 (free).
• Session token tracking with context window percentage display (136.4K (68%)) in chat footer.
• maxOutputTokens cap (8192) for paid OpenRouter models to prevent 402 errors.
• Render cold start error message improved — 'server was inactive' instead of 'cannot reach server'.
• Human-readable token formatting (5.2K, 136.4K, 1.0M).

Docs

• Rewamp authentication docs covering CLI device code flow, token storage, GitHub OAuth, and Render cold start troubleshooting.
• Rewamp terminal architecture docs reflecting actual codebase (CLI commands, DB models, API endpoints, AI providers, deployment).