Developers
Build with Scholar Feed
Search 600,000+ CS/AI research papers from your AI editor. No signup required.
Get started in 30 seconds
Scholar Feed is available as an MCP server that works with Claude Code, Cursor, VS Code, Windsurf, Zed, JetBrains, and any MCP-compatible client. Pick yours:
claude mcp add scholar-feed -- npx -y scholar-feed-mcpRestart Claude Code, then ask it to search for papers.
Any other MCP client works too — use the standard mcpServers block. No API key required (100 calls/day). For 1,000/day, add "env": { "SF_API_KEY": "sf_…" } (or -e SF_API_KEY=sf_… for Claude Code) and get a free key.
That's it. Try asking: "Search for recent papers on test-time compute scaling"
Prefer a guided setup? Run npx scholar-feed-mcp init to auto-detect your client and write the config.
What you can do
Search & discover
Full-text and semantic search across 600k+ papers with filters for novelty, category, recency, and method.
Compare methods
Side-by-side model comparisons — pull reported scores from paper full-text via the /compare-methods skill.
Author intelligence
Find researchers by topic or name, browse their papers, and see h-index, citation stats, and co-author graphs.
Paper content
Extract results from LaTeX source and export BibTeX for any paper via get_paper(format='bibtex').
23 tools, zero config
Every tool works out of the box. No API key needed for read & search. Just install and start asking questions.
Usage tiers
Start without an account. Upgrade when you need more.
| Tier | Daily limit | Features |
|---|---|---|
| Anonymous | 100/day | Read & search tools (no key required) |
| Free account | 1,000/day | All tools (shared across your keys) · 1 Field Guide + 1 ask-my-library / month |
| Pro / trial | 10,000/day | Everything, plus Pro-only gaps & embeddings · 50 Field Guides + 200 library answers / day |
Get a free API key for higher rate limits. Need more? Request a higher limit.
Example
Ask your AI assistant to search for papers:
You
"Find recent papers on chain-of-thought reasoning with novelty score above 0.5"
Tool call
search_papers(q="chain-of-thought reasoning", novelty_min=0.5, days=30)Result
Papers with titles, LLM summaries, novelty scores, citation counts, and links.