Bolt.diy
The open-source AI app builder: prompt, run, and deploy web apps with any LLM you choose.
About
Bolt.diy is the community open-source version of Bolt.new, started by Cole Medin. It brings the same in-browser AI app-building experience — describe an app, watch the agent write the code, run it, and deploy — but you choose the language model for every prompt, from OpenAI and Anthropic to local models via Ollama.
The project is MIT-licensed on GitHub with an active contributor community and official docs covering local, Docker, and Cloudflare Pages deployment. The software itself is free; you pay only for your own LLM API usage, which makes it the cost-conscious alternative to hosted builders, at the price of setting up and hosting it yourself.
What it can do
- In-browser AI agent that writes full-stack web apps from prompts
- Choice of 19+ LLM providers per prompt, including local models
- Live preview with run/edit/iterate loop
- Supabase integration for backend features
- One-click deploy to Netlify, Vercel, GitHub Pages
- Docker and Cloudflare Pages self-hosting guides
- MIT-licensed, community-maintained codebase
Setup
- Install Node.js (or Docker) on your machine.
- Clone the bolt.diy repository from GitHub.
- Install dependencies and start the app locally (see the official docs for OS-specific steps).
- Add your own LLM API key(s) in the settings (OpenAI, Anthropic, Ollama, OpenRouter, and others supported).
- Prompt the agent to build your app, preview it in the browser, and deploy to Netlify, Vercel, GitHub Pages, or Cloudflare Pages.
Prerequisites
Node.js or Docker; your own LLM provider API key(s) (or a local model via Ollama); a hosting target for deployment.
Use cases
Works with
Trust score — 7/10
Scored 7/10 (TIER verified, metrics-v1). Evidence: MIT-licensed open-source repo under coleam00 (bolt.new-any-llm, now bolt.diy) with documented 19+ LLM provider support and official docs at stackblitz-labs.github.io/bolt.diy; independent Medium comparison corroborates the feature and cost-model differences vs Bolt.new. Docs corroboration: README documents setup, provider configuration, and deployment paths. Strengths: fully open source (no lock-in); bring-your-own-model cost control; active community. Docked for: (1) requires meaningful technical setup (Node/Docker, API keys) — not no-code; (2) community project, so support and longevity depend on volunteer maintenance. License: MIT.
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 7.0/10
- Maintenance · recency, releases, issue load
- 7.0/10
- Openness · license + docs
- 8.0/10
- Manual review · sanity read + red-flag check
- 7.0/10
- · MIT-licensed open-source repo (coleam00/bolt.new-any-llm, now bolt.diy), started by Cole Medin, community-maintained
- · Documented 19+ LLM provider support incl. local models; official docs at stackblitz-labs.github.io/bolt.diy
- · Independent Medium comparison corroborates feature/cost differences vs Bolt.new; software free, user pays own LLM API costs
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.