Documentation
BuildAI generates React + TypeScript UIs with Tailwind, previewed on this server (no Tailwind CDN). Use the builder chat to plan, then confirm to generate code — or quick-build without a plan.
What this page is for
Docs is the place for environment setup, model configuration, integrations (e.g. Supabase), export/download behaviour, and limits. Product marketing lives on Pricing; the live product is the Builder.
AI models (header: Model)
- Auto (env) — Uses server rules: checks
AI_PROVIDER, then available keys. Typical order: Gemini (API key or Vertex) if configured, otherwise OpenAI ifOPENAI_API_KEYis set. This matches what most teams run in production (fast iteration + regional Vertex options). - Gemini — Good for codegen and JSON; set
GOOGLE_API_KEY/GEMINI_API_KEYor Vertex service account /GOOGLE_APPLICATION_CREDENTIALS. - Claude (Vertex) — Strong reasoning; requires GCP auth (not-consumer API key alone). Set
AI_PROVIDER=vertex_claudefor server default or pick it in the UI. - OpenAI — Set
OPENAI_API_KEY. Useful when Google APIs are region-restricted.
Copy .env.example to .env.local and add your keys. Restart yarn dev after changes.
UI pattern (chat: UI pattern)
Pick a preset so planning and codegen reuse the same shell: Admin shell (sidebar + top bar) or Storefront (shop nav + grid + cart), or Storefront + admin (one app: storefront + role-gated admin CRUD), SaaS marketing (landing sections), or Dashboard analytics (KPI + filters + table). Default stays flexible if you do not need a fixed framework.
Preview & export
Preview bundles via POST /api/preview-bundle (esbuild + Tailwind v4). Use Download in the preview toolbar to export a ZIP of generated sources.