bolt.diy - AI Code Assistants Tool

Overview

bolt.diy is an open-source AI coding assistant (MIT-licensed) that lets developers prompt, run, edit and deploy full‑stack JavaScript/Node.js web applications while choosing the LLM per prompt. It provides an integrated development experience that combines a WebContainer-backed live preview and terminal, Git integration, one‑click deployment targets (Vercel, Netlify, GitHub Pages), Supabase backend support, and an Electron desktop build so you can run the same tooling locally or in the browser. ([github.com](https://github.com/stackblitz-labs/bolt.diy)) The project exposes a modular provider architecture (BaseProvider) so teams can add cloud or local LLMs (OpenAI, Anthropic, Google/Gemini, Ollama, Mistral, Groq, HuggingFace, OpenRouter, Amazon Bedrock, and more) and supports MCP (Model Context Protocol) servers to enable tool-calling (database queries, file ops, external APIs) during AI-assisted workflows. bolt.diy targets developers who want an extensible, end-to-end AI developer loop (prompt → code edit → run → deploy) with live diffs, file-locking, and project snapshots. Note: although bolt.diy is MIT-licensed, its WebContainer runtime relies on a WebContainers API with separate commercial licensing requirements for production commercial use. ([stackblitz-labs.github.io](https://stackblitz-labs.github.io/bolt.diy/))

GitHub Statistics

• Stars: 19,071
• Forks: 10,355
• Contributors: 112
• License: MIT
• Primary Language: TypeScript
• Last Updated: 2026-02-07T14:36:22Z
• Latest Release: v1.0.0

The repository is actively maintained and broadly adopted: the project page shows ~19k stars and ~14k forks, with an active issues/PR backlog (dozens of open issues and pull requests), indicating a large user base and active community contribution. Recent commit history shows continued development into 2026 (example: commits on Feb 7, 2026). The codebase includes many examples, docs, and scripts (Docker, Electron, pnpm/npm scripts), and the project is packaged as a public template to encourage forks and contributions. These signals point to strong interest and ongoing maintenance, though users should review open issues and the project’s troubleshooting notes before adopting for production. ([github.com](https://github.com/stackblitz-labs/bolt.diy))

Installation

Install via npm:

git clone https://github.com/stackblitz-labs/bolt.diy

cd bolt.diy

cp .env.example .env.local

cp .env.local .env

sudo npm install -g pnpm  # install pnpm if you don’t have it

pnpm install

pnpm run dev  # start dev server (Remix + Vite)

# Docker (optional):

npm run dockerbuild  # build development image (or: docker build . --target bolt-ai-development)

docker compose --profile development up  # run container with hot reload

# Desktop (Electron): download prebuilt binaries from Releases or build locally via:

pnpm electron:build:dist  # build Electron app for all platforms

Key Features

• Per-prompt LLM selection across 19+ providers (OpenAI, Anthropic, Gemini, Ollama, Mistral, Groq, etc.).
• WebContainer live preview with integrated terminal to run and test Node.js apps in-browser.
• MCP (Model Context Protocol) support to let the AI call external tools, DBs, and APIs.
• Git integration with automatic commits, diff visualization, snapshots, and GitHub export.
• One-click deploy to Vercel, Netlify, or GitHub Pages from the project UI.

Community

bolt.diy has a large, active community (many stars and forks) and frequent commits and PRs; documentation and official docs site provide comprehensive setup and extension guides. Community feedback shows both strong enthusiasm and practical concerns: users praise the feature set and extensibility but report stability issues around WebContainer limits, long-running edits, and token consumption in some cases (see community threads and troubleshooting docs). The project maintains an open discussion hub and community resources (oTTomator Think Tank) and accepts contributions via GitHub. ([github.com](https://github.com/stackblitz-labs/bolt.diy))

Key Information

• Category: Code Assistants
• Type: AI Code Assistants Tool