Huabu

Where you and your agents think together.

User Handbook · Download · Feedback

Huabu gives you an infinite work surface where you and your agents think together. Bring ideas, files, open questions, and agents into one living surface, where they stay visible and connected across sessions—and turn clearer thinking into work that ships.

• Bring scattered information together — drag ideas, notes, documents, and outputs from different chat sessions and sources into one Space, so everything relevant stays visible and connected.
• Organize your thinking with AI — work with AI to arrange information, uncover relationships, and clarify ideas while reviewing and controlling every proposed change.
• Let your agents work with richer context — bring your own agents into a Space, where they can understand not only the materials you provide, but also how those materials relate to one another—helping them better grasp your intent and do more relevant work.

Packaged desktop applications are published through GitHub Releases. You can also build from source.

1. Download the latest package for your platform from the Releases page.
2. On macOS, open the .dmg; on Windows, run the .exe installer.
3. Launch Huabu and choose a local folder as your Home.

The current desktop packages target macOS on Apple silicon and Windows on x64. Available packages may vary by release; consult the release notes before installing.

Follow the Huabu User Handbook for setup instructions, model configuration, core concepts, and guidance on creating and working in your first Space. The handbook is the canonical source for product usage instructions.

Huabu does not provide an LLM service as part of the application. Use of a model or external capability may require a separate account, credentials, subscription, or usage charges from its provider.

Requirements: Node.js 20+ and pnpm 10+.

pnpm install

The source development server needs a stable encryption key before credentials can be saved through the Settings UI. Copy .env.example to .env, generate a key with the following command, and paste its output into HUABU_SECRET_KEY=:

node -e "console.log(require('node:crypto').randomBytes(32).toString('base64'))"

Keep that key unchanged: existing encrypted credentials cannot be opened with a different key. The .env file is git-ignored. Packaged Electron releases use the operating system's secure storage and do not require HUABU_SECRET_KEY.

Then launch the desktop development environment:

pnpm run dev:desktop

This launches the desktop app (recommended), starting the server, the web client, and the shared package in watch mode, then opening Huabu in its own desktop window.

The repository ships opt-in git hooks that give you fast feedback before you commit or push. Enable them once per clone:

pnpm run hooks:install

This generates the hooks into your local .git/hooks directory (they are not tracked in the repository), wiring up:

• pre-commit — lint-staged (ESLint --fix + Prettier on staged files)
• pre-push — pnpm run typecheck

Skip a single run with --no-verify, or disable the hooks again with pnpm run hooks:uninstall. These hooks are purely a local convenience — the authoritative gate is CI (.github/workflows/ci.yml), which runs lint, format, and typecheck on every pull request regardless of local setup.

Configuring a model provider and connecting external coding agents work the same way as in a packaged build, so they are covered in the handbook: Models and capabilities and External agents. A source build reads its encryption key from HUABU_SECRET_KEY, whereas packaged releases use operating-system-protected storage.

Huabu stores Spaces and their materials in the Home folder selected by the user. Credentials saved in the packaged desktop application are encrypted at rest using operating-system-protected storage.

Huabu does not send telemetry, crash reports, diagnostic logs, or usage data to Microsoft.

When an AI feature is used, the configured model provider or connected service may receive the prompt and relevant Space content needed to fulfill the request. External agents may also read or modify files within the working directory configured for them. Review the terms, privacy practices, data-handling controls, and permissions of every provider or agent before connecting it, and do not provide sensitive material unless its use is authorized.

Huabu is released for research and experimental use. AI-generated output may be inaccurate, incomplete, biased, or unsafe, and Huabu has not been evaluated for every language, model, or downstream scenario. Do not use it for high-risk or time-critical decisions, or in workflows where output is acted upon without qualified human review.

For the complete intended-use statement, evaluation summary, limitations, and responsible-use guidance, read RAI_README.md.

Community support is available on a best-effort basis through the Huabu issue tracker. See SUPPORT.md for support guidance. Report security vulnerabilities privately by following SECURITY.md.

We welcome bug reports, documentation feedback, and research feedback. See CONTRIBUTING.md for contribution guidance and Contributor License Agreement requirements. This project follows the Microsoft Open Source Code of Conduct.

This project may contain trademarks or logos for projects, products, or services. Authorized use of Microsoft trademarks or logos is subject to and must follow Microsoft's Trademark & Brand Guidelines. Use of Microsoft trademarks or logos in modified versions of this project must not cause confusion or imply Microsoft sponsorship. Any use of third-party trademarks or logos is subject to those third parties' policies.

The materials included directly in this repository are licensed under the MIT License, except where otherwise noted. The packaged Huabu application distributed through GitHub Releases is provided under the license terms included with that release.