Semantic Scholar alternative (MCP)
Scholar Feed: a Semantic Scholar alternative your AI assistant can call directly
Scholar Feed is a focused alternative to the Semantic Scholar API for CS/AI/ML, built as an MCP server your assistant calls with no glue code. Semantic Scholar is the better choice for breadth — 200M+ papers across every field, with a free REST API for building your own app. Scholar Feed trades that breadth for a curated 600,000+ CS/AI/ML corpus, an LLM novelty score on every paper, daily "watches" for new work, and zero-code use inside Claude Code or Cursor. Install with npx scholar-feed-mcp init.
Why people search "Semantic Scholar alternative for MCP"
Semantic Scholar (from the Allen Institute for AI) is the default free citation graph for a lot of researchers, and its Academic Graph API is genuinely excellent. Two things send people looking for an MCP-shaped alternative:
- The API is something you wire up in code. If you just want your AI assistant to search papers mid-conversation, you either write a wrapper or install someone else’s. An MCP server is that, out of the box.
- Coverage is broad but general. If you live in CS/AI/ML you want ranking and signals tuned to that firehose (novelty, code availability, citation velocity) rather than an all-fields graph.
Note: thin MCP wrappers over the Semantic Scholar API do exist. Scholar Feed isn’t one of those — it’s a separate corpus and ranking, described below.
How Scholar Feed compares
| Comparison axis | Semantic Scholar | Scholar Feed |
|---|---|---|
| Access shape | REST API (write code) or website | MCP server (assistant calls it directly, no code) |
| Coverage | 200M+ papers, all fields | 600,000+ CS/AI/ML papers, indexed daily from arXiv |
| Per-paper signal | TLDR summary | LLM summary + 0–1 novelty score |
| Keep-up | Email alerts (follow authors/papers on the site) | Daily watches on a saved filter (lab, technique, author, citation scope) |
| Citation graph | Authoritative, very large | Both directions, scoped to the CS/ML corpus |
| Full text | Some, via API | Extracts results/experiments from LaTeX source |
| Ranking | General relevance | Multi-signal (recency, citation velocity, code, institution), tuned for CS/ML |
| Build-your-own app | Free API is purpose-built for it | Not a general data API; it’s an assistant tool |
What you actually do with it
You ask your assistant, in plain language, "find recent high-novelty work on test-time compute scaling," and get ranked CS/ML papers with summaries in the window you’re already working in — no API calls to write. Then "set a watch on new retrieval-augmented-generation papers above 0.5 novelty" and it surfaces matches daily. The Semantic Scholar instinct (a real citation graph behind your search), delivered as a tool your agent uses, with a novelty filter to skip the incremental flood.
When NOT to use Scholar Feed
- You work outside CS/AI/ML, or you need the full 200M-paper graph. Semantic Scholar’s breadth and authoritative citation graph win clearly; Scholar Feed is a 600k CS/AI/ML corpus.
- You’re building an application and want a free, general-purpose data API. The Semantic Scholar Academic Graph API is built for that. Scholar Feed is an assistant tool (MCP), not a data backend.
- You specifically want a thin MCP wrapper over Semantic Scholar’s own data. Those exist; Scholar Feed is a different corpus and ranking, not an S2 proxy.
Frequently asked questions
Is Scholar Feed a good Semantic Scholar alternative for MCP?
Yes, if you want a CS/AI/ML literature tool your AI assistant can call directly. Unlike a REST API you wire up in code, Scholar Feed is an MCP server your assistant uses with no glue code. It adds an LLM summary and a 0 to 1 novelty score to each paper and extracts full text from LaTeX source. For all-fields breadth or the full 200M-paper citation graph, Semantic Scholar is still the better choice.
Is Scholar Feed just an MCP wrapper around the Semantic Scholar API?
No. Thin MCP wrappers over the Semantic Scholar API do exist, but Scholar Feed is not one of them. It is a separate corpus of 600,000+ CS/AI/ML papers indexed daily from arXiv, with its own multi-signal ranking and novelty scoring, not a proxy for Semantic Scholar’s data.
Do I need an account or API key to try it?
No. The search and read tools work anonymously at 100 calls per day. A free API key raises the limit to 1,000 calls per day, and Pro raises it to 10,000. Install with npx scholar-feed-mcp init.
Try it
npx scholar-feed-mcp initFree anonymous access is 100 calls/day (no account); a free key raises it to 1,000/day. Open source (MIT): scholar-feed-mcp on GitHub.
More setup options on the developers page.