Bolt.diy
A self-hosted AI coding assistant for full-stack web apps, with Ollama and LM Studio support, desktop apps, and MIT-licensed source code.
Bolt.diy is a self-hosted AI coding assistant for people building full-stack Node.js web apps who want to choose their own model provider. It's a fork of Bolt.new, with a browser workspace and an Electron desktop app for macOS, Windows, and Linux. You can also run it through Docker.
You choose the model. Ollama and LM Studio let it use models running locally, while OpenAI-compatible endpoints support other backends. Cloud choices include Anthropic, OpenAI, Gemini, and OpenRouter. The app can run on your machine, but choosing a cloud provider sends model requests to that service and requires its API credentials.
The workspace combines AI code generation with editing, an integrated terminal, and codebase search. Image attachments and voice prompts give you alternatives to text input. A diff view shows the AI's edits, file locking protects selected files during generation, and version restoration lets you return to earlier code. Git import, ZIP downloads, and host-folder syncing let you work with projects outside the app.
Supabase integration covers database management and queries, and MCP connects the assistant to additional tools. It also supports Expo projects for React Native apps and deployment to Netlify, Vercel, or GitHub Pages. The source is open source under MIT, but its WebContainer API requires a commercial license for production use in a for-profit setting; prototypes and proofs of concept don't require that license.
