Build OpenClaw from its open-source Git repository, write and register custom hardware plugins, and submit your first pull request to the OpenClaw open-source project. Covers CMake configuration, dependency management, testing workflows, and the contribution process.
Required: CMake 3.22+, GCC 11+ or Clang 14+, Python 3.10+, Git 2.35+, and 4 GB of free disk space. Optional but recommended: NVIDIA CUDA 12.0+ for GPU simulation, Eigen 3.4 for optimised matrix operations, and the Bullet Physics development headers (libbullet-dev on Ubuntu). Install all Ubuntu prerequisites: sudo apt install build-essential cmake git python3.11 python3.11-dev libeigen3-dev libbullet-dev libssl-dev.
Clone the repository: git clone https://github.com/openclaw/openclaw.git && cd openclaw. Create a build directory: mkdir build && cd build. Configure CMake: cmake .. -DCMAKE_BUILD_TYPE=Release -DOPENCLAW_BUILD_TESTS=ON -DOPENCLAW_BUILD_SIMULATOR=ON. Build: cmake --build . --parallel $(nproc). Install system-wide: sudo cmake --install . --prefix /usr/local. Verify: openclaw --version should return the version matching the Git tag you built.
Use the plugin template: openclaw plugin new --name my_gripper --type hardware. This generates a CMakeLists.txt, my_gripper_driver.cpp, and my_gripper_driver.h in ./my_gripper/. Implement the IHardwareDriver interface: override initialize(), move_to_position(), get_joint_states(), and emergency_stop(). Build: mkdir build && cmake .. -DOPENCLAW_DIR=/usr/local/share/openclaw/cmake && cmake --build . Install: sudo cmake --install . Verify: openclaw plugin list should show my_gripper.
From the build directory, run the full test suite: ctest --output-on-failure. Run specific test groups: ctest -R plugin_api for plugin tests, ctest -R hardware for hardware driver tests, ctest -R simulator for physics simulation tests. Generate a test coverage report (requires lcov): cmake .. -DOPENCLAW_ENABLE_COVERAGE=ON && cmake --build . && ctest && genhtml coverage.info --output-directory coverage_html.
Fork the openclaw/openclaw repository on GitHub. Create a feature branch: git checkout -b add-myrobot-driver. Implement your driver in src/drivers/myrobot/ following the IHardwareDriver interface. Add tests in test/drivers/myrobot_test.cpp. Run linting: clang-tidy src/drivers/myrobot/*.cpp and clang-format --dry-run src/drivers/myrobot/*.cpp. All tests and linting must pass locally before opening a PR. Submit a pull request with the driver name, supported hardware models, and a link to the hardware API documentation.
Run cmake with verbose output: cmake --build . --parallel 1 -- VERBOSE=1 2>&1 | tee build.log. Search build.log for "Error" and "fatal error". Common issues: missing libbullet-dev (install with sudo apt install libbullet-dev); CMake version too old (update with sudo snap install cmake --classic); Python.h not found (install python3.11-dev). Run openclaw doctor after successful build to verify the installation is correctly configured.
OpenClaw follows the Google C++ Style Guide. All new code must include unit tests with >80% line coverage. Run clang-format -i src/**/*.cpp src/**/*.h to auto-format before committing. Run clang-tidy src/**/*.cpp to check for style violations. Sign commits with your GPG key if contributing to the core framework. The project uses GitHub Actions for CI; all CI checks must pass before maintainers review a PR. For large contributions, open a GitHub Issue first to discuss the design before writing code.
Integrate OpenAI GPT-4o, Anthropic Claude 3.5 Sonnet, Google Gemini 2.0 Flash, and self-hosted Ollama LLaMA 3 with OpenClaw for natural-language robot instruction and AI-powered manipulation planning. Includes detailed per-call and monthly cost breakdowns for each AI provider.
Deploy OpenClaw on AWS EC2 and EKS, Azure Virtual Machines and AKS, and Google Cloud Compute Engine and GKE. Includes detailed cost breakdowns per cloud, GPU instance sizing for simulation, and architecture patterns for hybrid cloud-to-robot connectivity.
Deploy OpenClaw inside Docker containers for reproducible, portable robotics development. This guide covers single-container setups, multi-service docker-compose stacks, GPU passthrough for simulation, and production image hardening.