git-copilot is an AI-powered code review CLI that runs locally and supports 40+ LLM providers.
# Using npm
npm install -g git-copilot
# Using Homebrew (if published)
brew install yourusername/tap/git-copilot
# Using pre-built binary (from GitHub Releases)
curl -L https://github.com/yourusername/git-copilot/releases/download/v1.0.0/git-copilot-linux-x64 -o git-copilot && chmod +x git-copilot && sudo mv git-copilot /usr/local/bin/- Node.js 20+ (if installing via npm)
- A supported LLM provider API key (OpenAI, Anthropic, Ollama, etc.)
- For Beads integration:
bdCLI (optional, install from steveyegge/beads)
- Configure your LLM provider:
git-copilot config set providers.openai.apiKey sk-...
# or for anthropic:
git-copilot config set providers.anthropic.apiKey sk-ant-...
# set active provider:
git-copilot config set activeProvider openai- (Optional) Initialize Beads for task coordination:
bd initgit-copilot reviewThis will:
- Analyze all changed files in the current branch
- Run parallel agents (security, performance, architecture, etc.)
- Show a progress dashboard in the terminal
- Output a comprehensive report
git-copilot review --range HEAD~5..HEADgit-copilot review --format markdown --output report.md
git-copilot review --format html --output report.html
git-copilot review --format pdf --output report.pdf
git-copilot review --format docx --output report.docx
git-copilot review --format pptx --output report.pptx
git-copilot review --format xlsx --output report.xlsxgit-copilot graphUse j/k to navigate, Enter to view commit details, / to search.
git-copilot dashboardgit-copilot stores configuration in ~/.git-copilot/config.yaml (or use the CLI to set values).
Common settings:
activeProvider: openai
providers:
openai:
apiKey: sk-...
model: gpt-4o
anthropic:
apiKey: sk-ant-...
model: claude-3-5-sonnet-20241022
maxConcurrent: 4
output:
format: terminal
beads:
external:
enabled: true
cliPath: bd
dataDir: .beads
custom:
enabled: true
retentionDays: 90Use git-copilot config get <key> and git-copilot config set <key> <value> to manage config.
git-copilot runs multiple specialized agents in parallel:
- SecurityAgent: OWASP Top 10, XSS, SQL injection, etc.
- PerformanceAgent: Complexity, N+1 queries, memory leaks
- ArchitectureAgent: Design patterns, coupling, cohesion
- CodeQualityAgent: Complexity, duplication, readability
- DependencyAgent: CVE checks, outdated dependencies
- GitHistoryAgent: Commit hygiene, secrets detection
- AggregatorAgent: Deduplication and severity scoring
- ReportWriterAgent: Formats final report
Each agent's findings are stored in the Beads memory system for cross-session context.
When using Anthropic Claude, git-copilot automatically loads relevant Skills for domain-specific expertise. Built-in Skills include:
code-review- General code review patternssecure-code-review- Security-focused reviewowasp-audit- OWASP complianceweb-design-audit- Frontend best practicesdocx,pdf,pptx,xlsx- Format-specific output rules- And more...
Skills are loaded from the skills/ directory and can be customized.
Beads provides task coordination and memory:
- External Beads (
bdCLI) tracks review tasks, dependencies, and status - Custom Memory (SQLite) stores findings for semantic search and context injection
Enable/disable in config under beads section.
Ensure you're using Node.js 20+ and run npm ci to install exact dependencies.
Check your API key and network connectivity. For Ollama, ensure the server is running locally.
Install Beads: npm install -g @beads/bd or follow instructions at https://github.com/steveyegge/beads
Add your user to the appropriate group or use sudo for global install.
- Read the full Configuration reference
- Learn about Agents and what each one does
- Explore Skills and how to create custom ones
- Understand Beads setup and troubleshooting
- Optimize Performance for large repos
- Check Troubleshooting for common issues