Llama Coder at a glance
Checked 16 Oct 2026Llama Coder is a Next.js web app that turns one prompt (or a screenshot) into a small single-page React app and previews it in a sandboxed browser iframe. Models are open-weight ones served by Together AI (recent commits default to DeepSeek V4.1 Flash and GLM); there is no local-model option. Self-hosting needs a Together AI key, a Postgres database (Neon suggested) and S3 for uploads. It does not build backends, databases or deploy apps. MIT; six contributors; active in October 2026.
RepoQuarry. “Llama Coder.” Checked 16 Oct 2026. https://repoquarry.dev/projects/llamacoderHow much of Lovable does it replace?
3 of 6 jobs. Small one-file React apps and prototypes only; no backend, database, editing canvas or publishing.
For change the UI by clicking on it (Visual Edits), add a database, publish the app to a live URL (Publish), see other Lovable alternatives. You may also need to build an interface, an API server and hosting yourself.
Facts, with sources
Every answer links to where we found it. We read the documentation; we did not install it.
Run `npm install` and `npm run dev` to install dependencies and run locallygithub.com · llamacoder#cloning--running ↗ 1d ago
Powered by Llama 3 on Together.ai.github.com · llamacoder#cloning--running ↗ 1d ago
Create a `.env` file and add your API keys:github.com · llamacoder#cloning--running ↗ 1d ago