447a3d66c0
- Introduced a minimal documentation chatbot that builds a search index from markdown files and serves responses via an API. - Added scripts for building the index and serving the chat API. - Updated package.json with new commands for chat index building and serving. - Created a new Vercel configuration file for deployment. - Added a README for the docs chat prototype detailing usage and integration.
Docs Chat Prototype
Minimal docs chatbot that reads source docs, builds a section index, and answers questions from those excerpts via OpenAI.
Build the index
pnpm docs:chat:index
This generates scripts/docs-chat/search-index.json from docs/**/*.md.
Pipeline Integration
The docs-chat context is the generated index. CI rebuilds it whenever docs change
so PRs keep scripts/docs-chat/search-index.json in sync. If you run docs
publishing outside CI (for example via Mintlify), make sure the deploy pipeline
also runs pnpm docs:chat:index so the chat context stays current.
Run the API
OPENAI_API_KEY=sk-... pnpm docs:chat:serve
Defaults to http://localhost:3001. Health check:
curl http://localhost:3001/health
Mintlify widget
Mintlify loads any .js in the docs content directory on every page.
docs/assets/docs-chat-widget.js injects a floating “Ask Molty 🦞" button and
calls the API at:
window.DOCS_CHAT_API_URL || "http://localhost:3001"
To use a deployed API, set window.DOCS_CHAT_API_URL before the widget runs
(for example by adding another small .js file in docs/assets/ that sets it).