- 1. MCP Support
- See openai#270 and https://docs.anthropic.com/en/docs/agents-and-tools/claude-code/tutorials#set-up-model-context-protocol-mcp
- This may take care of items 4 5, and 8.
- 2. Local Model Support
- See openai#26
- 3. Cost Management
- Cost tracking in progress: openai#117
- See discussion on rate limit: openai#328
- 4. Web Search Integration
- See openai#259
- 5. Browser/Desktop Agent
- 6. Voice Input
- See openai#418
- 7. Bash History Integration
- See openai#350 (kinda)
- 8. Memory
- See openai#323 and openai#587 and openai#341
From npm (Recommended)
TODO
Build from source
# Clone the repository
git clone https://github.com/huyouare/codex.git
cd codex
# Enable corepack
corepack enable
# Install dependencies and build
pnpm install
pnpm build
# Run the locally‑built CLI directly
node ./dist/cli.js
# Or link the command globally
pnpm link
We welcome contributions! Here's how to get started:
- Fork and clone the repository
- Install dependencies:
pnpm install
- Enable Git hooks:
pnpm prepare
# Watch mode (tests rerun on change)
pnpm test:watch
# Type‑check without emitting files
pnpm typecheck
# Automatically fix lint + prettier issues
pnpm lint:fix
pnpm format:fix
We use Husky to enforce code quality:
- Pre-commit: Runs lint-staged to format and lint files
- Pre-push: Runs tests and type checking
- Create a topic branch from
main
(e.g.,feat/web-search
) - Make your changes and add tests
- Run all checks locally:
pnpm test && pnpm lint && pnpm typecheck
- Push and open a PR
- Fill in the PR template explaining What/Why/How
- Ensure all checks pass
- Node.js 22 or newer (LTS recommended)
- macOS 12+, Ubuntu 20.04+/Debian 10+, or Windows 11 via WSL2
- Git (recommended)
- 4GB RAM minimum (8GB recommended)
Apache-2.0 License