docs-search
Search indexed documentation for relevant information using mem-oracle.
When to Use
Use this skill when the user asks about:
- •How to use a specific library, framework, or API
- •Documentation for a package or tool
- •Best practices from official docs
- •Code examples from documentation
- •Configuration or setup guides
How to Use
- •First check if the worker is running by making a health check
- •Use the retrieve endpoint to search for relevant documentation
- •Present the results with source URLs
API Endpoints
Base URL: http://127.0.0.1:7432
Health Check
bash
curl http://127.0.0.1:7432/health
Search Documentation
bash
curl -X POST http://127.0.0.1:7432/retrieve \
-H "Content-Type: application/json" \
-d '{"query": "your search query", "topK": 5}'
Index New Documentation
To index a new documentation site:
bash
curl -X POST http://127.0.0.1:7432/index \
-H "Content-Type: application/json" \
-d '{"baseUrl": "https://docs.example.com", "seedSlug": "/getting-started"}'
Check Indexing Status
bash
curl http://127.0.0.1:7432/status
Example Queries
- •"How do I use server components in Next.js?"
- •"What are the best practices for React hooks?"
- •"How to configure TypeScript paths?"
- •"Explain Tailwind CSS dark mode setup"
Response Format
Results include:
- •
title: Page title - •
heading: Section heading (if available) - •
content: Relevant text snippet - •
url: Source documentation URL - •
score: Relevance score (0-1)
Indexing Documentation URLs
When you encounter a documentation URL in the conversation, you can index it:
bash
# Index Next.js docs
curl -X POST http://127.0.0.1:7432/index \
-H "Content-Type: application/json" \
-d '{"baseUrl": "https://nextjs.org", "seedSlug": "/docs/app"}'
# Index React docs
curl -X POST http://127.0.0.1:7432/index \
-H "Content-Type: application/json" \
-d '{"baseUrl": "https://react.dev", "seedSlug": "/reference"}'
Troubleshooting
If the worker is not responding:
- •Check if it's running:
curl http://127.0.0.1:7432/health - •View logs:
tail -f ~/.mem-oracle/worker.log - •Start manually:
cd /path/to/mem-oracle && bun run worker