Turn a docs site into agent context
Our verdict
Agents hallucinate APIs they don't know. Giving them the current docs as local Markdown is a simple fix. Refresh when the library version changes.
- Who it's for
- Builders integrating new or fast-moving APIs
- Leverage
- Fewer hallucinated API calls
- Effort
- Low
- Date added
What it does
Use a crawler (Firecrawl or Crawl4AI) to pull a docs site into Markdown files in your repo (e.g. /docs/vendor), or convert PDFs with MarkItDown, then reference them in your agent's instructions.
How builders use it
- Crawl the relevant docs section only, not the whole site.
- Reference the folder in CLAUDE.md / AGENTS.md.
- Ask the agent to cite the doc file it relied on.
What to expect
Integration code that matches the real, current API.
Copy this: Agent instruction
Before writing any code that calls <library>, read the relevant files in docs/vendor/<library>/.
Only use functions and parameters that appear there.
In your summary, list which doc files you used.Related discoveries
- Research & AutomationUse case
Weekly competitor digest with n8n + Firecrawl
Scrape competitor pages weekly, summarise changes with an LLM, email yourself.
Effort: Medium
- Coding & Shipping AgentsUse case
Give your agent a conventions file
A short CONVENTIONS.md stops agents from fighting your style every session.
Effort: Low
- Vibe-Coding & Prompt-to-AppTool
Context7: up-to-date docs for AI editors
Feeds current, version-specific library docs into your coding assistant.
Effort: Low
- Vibe-Coding & Prompt-to-AppUse case
Cursor Rules and AGENTS.md
Persistent instructions so the AI follows your stack every time.
Effort: Low
Want more like this?
Practical discoveries, without the noise.