A cognitive context engine for thoughtful engineers. Helps you understand what your code means and does, while you're writing it.

Fully standalone. No API server, no authentication, no billing. All analysis runs locally.

• 196 built-in rules across Python, Go, Rust, and TypeScript
• Impact analysis — "what breaks if I change this file?"
• Centrality analysis — find your most critical files
• RAG-powered queries — ask questions about your codebase
• LSP server — real-time diagnostics in your editor
• Auto-fix — apply suggested patches automatically

# Install
cargo install unfault

# Analyze your project
cd /path/to/your/project
unfault review

# Ask questions about your code
unfault ask "what are the main stability concerns?"

# Impact analysis
unfault graph impact src/auth.py

# Find critical files
unfault graph critical

# Start the LSP server (for IDE integration)
unfault lsp

Configuration is stored at ~/.config/unfault/config.json.

# Use Ollama (local, no API key needed)
unfault config llm ollama --endpoint http://localhost:11434 --model llama3.2

# Use OpenAI
unfault config llm openai --model gpt-4o

# Use Anthropic
unfault config llm anthropic --model claude-sonnet-4-20250514

Configure via ~/.config/unfault/config.json:

{
  "embeddings": {
    "provider": "ollama",
    "endpoint": "http://localhost:11434",
    "model": "nomic-embed-text",
    "dimensions": 768
  }
}

This is a Cargo workspace containing 4 publishable crates:

core ← analysis ← rag ← cli

cargo build --workspace

cargo test --workspace

Crates must be published in dependency order:

cargo publish -p unfault-core
cargo publish -p unfault-analysis
cargo publish -p unfault-rag
cargo publish -p unfault

MIT