// DOCUMENTATION

All documentation is hosted on GitHub alongside the source code.

// GETTING STARTED

README Project overview, quick start, and installation instructions ↗ User Guide Setup, configuration, and day-to-day usage ↗ Configuration ~/.finch/config.toml reference — providers, LoRA, model settings ↗ Changelog Version history and release notes ↗

// PROGRAM RUNTIME

Finch VM Contract Shared typed runtime, effects, capabilities, and provider wire contract ↗ Co-Forth Canonical concatenative syntax, definitions, stack effects, and examples ↗ Co-Lisp Readable prefix syntax that lowers to the same typed IR as Co-Forth ↗ Shared Program Runtime Architecture, durable vocabulary, and remaining runtime roadmap ↗

// ARCHITECTURE

CLAUDE.md Complete system architecture — models, TUI, tools, LoRA, providers ↗ Architecture System design overview and component diagram ↗ TUI Architecture Terminal UI rendering — scrollback, shadow buffer, diff-based blitting ↗ Daemon Mode HTTP server mode, OpenAI-compatible API, session management ↗ Model Backends ONNX vs Candle comparison, supported model families ↗

// FEATURES

Tool System Agentic tools, permission system, approved patterns ↗ LoRA Training Weighted feedback collection and background fine-tuning ↗ MCP Model Context Protocol support — connecting external tools ↗ Multi-Provider Configuring Claude, GPT-4, Gemini, Grok, Mistral, Groq, and local ↗

// DEVELOPMENT

Development Guide Building from source, running tests, contributing ↗ Roadmap Planned features, open issues, and future direction ↗ GitHub Issues Bug reports, feature requests, and open contributor tasks ↗ Pull Requests Open and merged contributions ↗