Who it is forSomeone building a research agent that needs current information.
The problemSearch APIs return only titles and snippets, so the agent still has to fetch each page.
The outcomeFull text of the top results for a query, ready to put in a prompt.
What you need first
- curl or any HTTP client for the hosted r.jina.ai and s.jina.ai endpoints
- Docker to self-host with `docker pull ghcr.io/jina-ai/reader:oss`
- An optional Jina API key for higher rate limits on the hosted service
Steps
- URL encode your query and prepend https://s.jina.ai/
- Restrict to sites with the site parameter, for example
curl 'https://s.jina.ai/When%20was%20Jina%20AI%20founded%3F?site=jina.ai&site=github.com' - Pass the returned markdown to your model as context
- Keep the source URLs from each result so answers can be checked
Prompt for your agent
With jina-ai/reader, query s.jina.ai for <my research question> limited to <two sites>, and summarize the results with their source URLs.
Caveat. The README says search fetches the top 5 results, so expect slower responses than a snippet API.
Use this in your AI agent
Ask your agent
Using the Deduped MCP, get the setup steps for jina-ai/reader and "Search the web and read the top results"
Add Deduped to Claude Code once
claude mcp add --transport http deduped https://dedupedai.com/mcp
Every Monday: the 5 hottest repos and exactly how to use them. Free, one email a week, unsubscribe anytime.
Get them freeShare: X LinkedIn Hacker News Email Copy link
Third-party repos we don't own. AI-written use cases can be wrong. Disclaimer