bolt.diy
Open-source AI coding workspace that builds full-stack web apps in the browser with your choice of LLM provider.
Prompt, edit and preview full-stack apps in the browser, with 22+ model providers to pick from.
bolt.diy is an open-source AI developer workspace that lets you pick the language model for each prompt. It suits developers who want a bolt.new-style app builder where they bring their own API keys or local models.
What it is
It runs a full-stack Node.js environment in the browser using WebContainers, with a code editor, terminal, visual diff viewer and live preview. You connect providers such as OpenAI, Anthropic, Gemini, DeepSeek, Ollama or LM Studio. Projects can be pushed to Git, exported as ZIP and deployed to Netlify, Vercel or GitHub Pages.
Who it's for
It suits developers and technical makers who are fine installing Node and pnpm or running Docker. It is a weaker fit for non-technical founders who want a hosted, ready-made builder, and the README says no new features are planned.
Key features
Provider choice per prompt
Pick among 22+ cloud and local LLM providers, including Ollama and LM Studio.
In-browser dev environment
WebContainers run a full-stack Node.js environment inside the browser.
Workbench with diff viewer
Editor, terminal, visual diffs and live preview sit in one workbench.
One-click deploys
Deploy to Netlify, Vercel or GitHub Pages from the app.
Git and snapshots
Clone and push Git repos, restore snapshots and export ZIP.
MCP and Supabase
Model Context Protocol tools and Supabase integration are built in.
Up and running in 3 steps
Our honest verdict
What's great
- You choose the model for each prompt, including local ones through Ollama or LM Studio.
- Git cloning and pushing, snapshot restore and ZIP export keep your work portable.
- It ships a desktop Electron app as well as the browser version.
- Built-in deploy to Netlify, Vercel and GitHub Pages shortens the path to a live preview.
Where it falls short
- Running it yourself needs Node.js 22.12 or newer and pnpm, or Docker, so it is not a one-click app for non-developers.
- You pay your LLM provider for API calls, or need hardware for local models.
- The README says ongoing work covers bug fixes and dependency updates, with no new features planned.
- API keys entered in the UI are stored in browser cookies, which you should weigh on shared machines.
Pricing & license
Free under the MIT licence with no paid plan shown in the README. You pay for any cloud LLM API you connect or for hardware to run local models.
Open remote AI-training roles
860 open remote AI-training roles, updated November 14, 2026; most pay $20–$170 an hour. Zekai lists them free; the application goes to the employer’s platform.
Alternatives
These reviewed AI tools do the same job instantly — from our scored directory.
Use it, or have us set it up
FAQ
Can I use local models?
Yes. The README lists Ollama and LM Studio as local providers, configured by URL in the UI or in .env.local.
Is there a desktop version?
Yes. The README offers pre-built Windows, macOS and Linux binaries from the latest release, or a build from source.
Does it have new features coming?
The README says ongoing work is bug fixes and dependency updates, and that no new features are planned.
