scholarly-metadata
Semantic Scholar (Academic Graph)
An academic search engine and citation graph from the Allen Institute for AI, covering about 220 million papers with citation context. Search on the website, or use the free Graph API, which needs no key but runs on a shared, heavily throttled anonymous pool. Especially handy for following citation trails and seeing how papers relate. Under load the keyless API is frequently rate-limited, so the website is the more dependable route.
Why it’s useful & how it works
Semantic Scholar is especially useful for following citation trails: it shows not just citation counts but also the sentence-level context in which papers cite each other, helping you quickly judge how a reference is actually used. The main website at semanticscholar.org is the most reliable way to search, since the API's keyless shared pool is frequently rate-limited under heavy load. You can search by title, author, or topic, and navigate forward and backward through citation networks. The website remains fully free and unpaywalled.
What’s inside
More than 200 million papers indexed across all academic fields, with over 2.4 billion citation edges linking them, plus records for more than 80 million authors.
API access
https://api.semanticscholar.org/graph/v1/paper/DOI: <doi>?fields=...
What we measured
Our own probes, not the archive’s own claims. Re-run periodically; every reading below is dated.
Reachability
- Direct request
- Responded HTTP 200 895 ms
- Through a datacenter proxy
- Responded HTTP 200 1.5 s
- API, direct
- Responded HTTP 200 762 ms
- API, through a proxy
- Responded HTTP 200 1.5 s
Did a real search return anything?
The request failed · median 693 ms · down during the run
Shared anonymous pool → frequent 429/transient error; needs backoff or an API key (no gmail keys).
Benchmark coverage
Found 1 of 9 test items (11%) in the Academic papers cluster. Median response 503 ms.
By how well known the item is: 0/3 well-known, 0/4 mid, 1/2 obscure .
One item it found: Quantum error correction below the surface code threshold
Reachability measured 2026-08-22. Search checked 2026-08-22. Benchmark run 2026-08-22.
Access
Programmatic API access (a key may be required, see the API tag).