SciSpace API
Scholarly search, document understanding, extraction and grounded answers over the SciSpace corpus.
Welcome
- IntroductionThe SciSpace API turns a research question into an answer you can check. Every sentence it returns carries a citation into the 280M+ works in the corpus — down to the page, and the
- What you can buildFive concrete products, each with the endpoints it needs and a link to the guide that builds it.
Get started
- QuickstartGet a cited answer to a research question in under five minutes.
- AuthenticationEvery request carries a secret API key as a bearer token over HTTPS. Requests without a valid key return 401 with error.type: authentication_error.
- API keysCreate, scope, rotate, and revoke keys from Settings → API keys.
- EnvironmentsTest and live keys address the same endpoints, the same corpus, and the same models. The difference is billing, throughput, and retention — not behaviour.
- Your first requestA slower walkthrough than quickstart, explaining every part of the call. Use this page if the quickstart worked but you want to know why.
- Errors and retriesEvery error returns the same envelope. Branch on error.code, never on message — messages change, codes are part of the compatibility contract.
- Client libraries
Concepts
- Data model overviewTwelve objects. Learn these four and the rest follow: Paper, Document, Search, Citation.
- PaperA paper is a scholarly work in the SciSpace corpus, identified by a pap_ ID and usually a DOI. You do not create papers; you look them up and reference them.
- DocumentA document is a PDF you uploaded, parsed and ready for questions and extraction. Private to your account, deletable, and metered on ingest.
- LibraryA library groups documents and carries access permissions. Every organization has a default library; creating more is optional.
- SearchA search is one literature query over the corpus, run at a chosen depth. It produces an Answer and the ranked papers that answer is grounded in.
- AnswerAn answer is generated prose plus the citations that support it. It is never a bare string.
- CitationA citation links a span of generated text to the source it came from — with the quoted evidence and, for PDFs, the exact region on the page. This is the object that makes output ch
- Extraction schemaAn extraction schema is the set of fields you want pulled out of every document — the column definitions of the table you are building.
- ChatA chat is a conversation scoped to a fixed set of documents and/or papers. The scope is set at creation and does not drift.
- MessageA message is one turn in a Chat. User turns carry text; assistant turns carry an Answer.
- JobA job is the async execution behind a resource. There is one status vocabulary across the whole API, so you write the wait logic once.
- WebhookA webhook delivers an event to your endpoint when an async job finishes, so you do not have to poll. Use them for anything that routinely takes more than 30 seconds — above all dee
- DepthDepth is how much effort to spend on a request. It is the only quality dial, and it never names a model — internal models can change without breaking your integration.
- CreditA credit is the unit of metering. Every metered response reports what it cost and what remains.
- Rate limitA rate limit caps throughput. It is separate from credits: credits limit how much you may spend, rate limits how fast you may spend it.
- TopicA topic is a concept extracted across the literature for a query, with the evidence that supports it — the ideas behind the papers rather than the papers themselves.
- FileA file is the upload primitive. You create a file, then turn it into a Document — and exports hand you a file back.
API reference
- Reference overviewBase URL https://api.scispace.com/v1. Read conventions once; it covers auth, pagination, idempotency, versioning, and headers so the resource pages do not repeat them.
- ConventionsEverything true of every endpoint. Resource pages link here instead of restating it.
- SearchesAnswer a research question over the corpus, with citations and the ranked papers behind the answer.
- PapersLook up scholarly metadata, full text where licensing allows, and citation graph edges.
- DocumentsIngest your own PDFs so you can chat with them and extract from them.
- ChatsAsk questions about a fixed set of documents and papers, and get answers with page-level citations.
- ExtractionsRun a field schema over many documents or papers and get a table back, with a citation per cell.
- ParaphrasesRewrite academic text with control over tone, length, and how far the wording may drift.
- AI detectionsAssess how likely a passage was machine-generated, sentence by sentence.
- TopicsExtract the concepts that matter for a query, each with the evidence behind it.
- CitationsFormat bibliography entries and in-text citations in over 2,000 journal styles.
- FilesUpload inputs and download exports.
- JobsOne endpoint to check the state of any async resource.
- Webhook endpointsRegister HTTPS URLs to receive completion events.
- UsageCredit balance and consumption, broken down by day, endpoint, and key. Free to call.
SDKs
Guides
- Build a literature review agentOutcome: a loop that takes a research question, searches at increasing depth, screens the papers, and returns a cited summary your users can verify.
- Extract data from a batch of PDFsOutcome: a CSV with one row per paper and a citation for every cell, built from a folder of PDFs.
- Ground an LLM answer in real citationsOutcome: answers in your product where every claim links to a source the user can open and check.
- Stream a long-running deep reviewOutcome: a deep review that shows progress in the UI instead of a spinner that times out.
- Render citation highlights in your own viewerOutcome: clicking a citation scrolls your PDF viewer to the exact sentence and draws a highlight over it.
- Sync results to a reference managerOutcome: search results land in Zotero, Mendeley, or EndNote as proper references.
- Migrate from OpenAlex or Semantic ScholarOutcome: your existing metadata pipeline keeps working, with grounded answers added on top.
- Control cost and creditsOutcome: predictable spend at scale, with no surprise 402s in production.
Integrations
- OverviewFour supported integration paths. Anything not listed here is a plain REST call — see reference-overview.
- LangChain and LlamaIndexUse SciSpace as a retriever so an existing RAG application gains scholarly grounding.
- Claude and OpenAI tool useExpose SciSpace endpoints as tools so an agent can research on demand.
- Zotero and MendeleyMove search results and bibliographies into a reference manager.
- Zapier
Platform
- Rate limits and quotasRequests are limited per API key. Exceeding a limit returns 429 with Retry-After in seconds.
- Pricing and creditsAPI usage is metered in credits. Every metered response reports its cost in X-Credits-Cost.
- Versioning and deprecationThe path pins the major version; a dated header pins behaviour within it.
- Content licensing and permitted useWhat you may do with the scholarly content this API returns.
- Data privacy and retentionWhat happens to the content you send us, and for how long.
- Security and compliance
- Reliability and status
- Teams, keys and scopesAccess has two layers: roles control what a person can do in the dashboard, scopes control what a key can do against the API.
Support
71 pages. Machine-readable index at /llms.txt; append
.md to any page URL for its Markdown source.