Skip to content
Scalekit Docs
Talk to an Engineer Dashboard

Connect AI agents to PDF Vector

Scalekit connector
Open markdown

The PDF Vector connector lets your AI agent act in each user's PDF Vector account. Each user connects their own PDF Vector access token once, and Scalekit sends it with every call, so your agent never handles credentials. It comes with 19 tools.

Tools
19
What they doRead · write · destructive
19 · 0 · 019 read0 write0 destructive
Users sign in with

Setup

  1. Install the SDK

    Terminal window
    npm install @scalekit-sdk/node dotenv
  2. Set your credentials

    Add your Scalekit credentials to your .env file. Find values in app.scalekit.com > Developers > API Credentials.

    .env
    SCALEKIT_ENVIRONMENT_URL=<your-environment-url>
    SCALEKIT_CLIENT_ID=<your-client-id>
    SCALEKIT_CLIENT_SECRET=<your-client-secret>
  3. Create the PDF Vector connection

    In AgentKit > Connections, create a PDF Vector connection. The name you give it is the connection_name your code passes. See Configure connections.

  4. Authorize a user and make your first call

    quickstart.mts
    import { ScalekitClient } from '@scalekit-sdk/node'
    import 'dotenv/config'
    import { createInterface } from 'node:readline/promises'
    const scalekit = new ScalekitClient(
    process.env.SCALEKIT_ENVIRONMENT_URL,
    process.env.SCALEKIT_CLIENT_ID,
    process.env.SCALEKIT_CLIENT_SECRET,
    )
    const actions = scalekit.actions
    const connector = 'pdfvector'
    const identifier = 'user_123'
    // Generate an authorization link for the user
    const { link } = await actions.getAuthorizationLink({ connectionName: connector, identifier })
    console.log('Authorize PDF Vector:', link)
    const rl = createInterface({ input: process.stdin, output: process.stdout })
    await rl.question('Press Enter after authorizing...')
    rl.close()
    // Make your first call
    const result = await actions.executeTool({
    connector,
    identifier,
    toolName: 'pdfvector_academic_parse',
    toolInput: {},
    })
    console.log(result)
    Terminal window
    npx tsx quickstart.mts

    Each user opens the link once and enters their PDF Vector credentials there. See Authorize a user for the full flow and statuses.

Tools

Pass the exact name to execute_tool
Try in PlaygroundRequest a tool
  • pdfvector_academic_fetchFetch academic publications by identifier: DOI, PubMed ID, arXiv ID, Semantic Scholar ID, ERIC ID, Europe PMC ID or OpenAlex ID.Read-only

    Fetch Academic Papers

    Fetch academic publications by identifier: DOI, PubMed ID, arXiv ID, Semantic Scholar ID, ERIC ID, Europe PMC ID or OpenAlex ID. Costs 2 credits. Returns each publication's metadata, such as title, authors, year, abstract, citation counts and PDF link. Use this to look up known papers. Use pdfvector_academic_search to find papers by topic.

    Inputs

    idsarrayrequired
    Publication identifiers to fetch; the provider is detected automatically. Example: ["10.48550/arXiv.1706.03762", "arXiv:2005.14165"]
    fieldsarray
    Publication fields to return. All fields except providerData are returned when empty. Example: ["title", "doi", "authors", "year", "abstract", "pdfURL"]
  • pdfvector_academic_find_citationsFind academic papers that support each sentence of a paragraph, using semantic similarity.Read-only

    Find Citations

    Find academic papers that support each sentence of a paragraph, using semantic similarity. Costs 2 credits per sentence. Returns, per sentence, the matching papers with relevance scores and metadata. Use this to add references to your own text. Use pdfvector_academic_search for a keyword search.

    Inputs

    paragraphstringrequired
    The text to find citations for. Example: Large language models can be adapted to new tasks with only a few examples provided in the prompt.
    fieldsarray
    Publication fields to return. All fields except providerData are returned when empty. Example: ["title", "doi", "authors", "year", "abstract", "pdfURL"]
    providersarray
    Databases to search. Defaults to semantic-scholar. Example: ["semantic-scholar", "openalex"]
  • pdfvector_academic_paper_graphGet the papers that cite a paper and the papers it references, as full publication objects.Read-only

    Get Paper Citations and References

    Get the papers that cite a paper and the papers it references, as full publication objects. Costs 2 or more credits, scaling with results. Returns citations and references lists with publication metadata and totals. Use this to explore a citation network. Use pdfvector_academic_similar_papers for related papers ranked by similarity.

    Inputs

    idstringrequired
    Paper identifier: DOI, Semantic Scholar, arXiv, PubMed or OpenAlex ID, or a URL. Example: 10.48550/arXiv.1706.03762.
    citations_limitinteger
    Maximum citing papers to return, 0 to 1000. Defaults to 100. Example: 50.
    citations_offsetinteger
    Citing papers to skip, for paging. Defaults to 0. Example: 0.
    fieldsarray
    Publication fields to return. All fields except providerData are returned when empty. Example: ["title", "doi", "authors", "year", "abstract", "pdfURL"]
    references_limitinteger
    Maximum referenced papers to return, 0 to 1000. Defaults to 100. Example: 50.
    references_offsetinteger
    Referenced papers to skip, for paging. Defaults to 0. Example: 0.
  • pdfvector_academic_parseFind the public PDF of an academic paper from its identifier or URL and convert it into markdown.Read-only

    Parse Academic Paper

    Find the public PDF of an academic paper from its identifier or URL and convert it into markdown. Returns the paper's markdown, page count and credits charged, or NOT_FOUND when no public PDF exists. Use this to read a paper's full text. Use pdfvector_document_parse for a PDF you already have a URL to.

    Inputs

    callback_urlstring
    Webhook URL for asynchronous processing. When set, PDF Vector returns 202 at once and POSTs the result to this URL. Example: https://example.com/hooks/pdfvector.
    idstring
    Publication identifier (DOI, PubMed, arXiv, Semantic Scholar, ERIC, Europe PMC or OpenAlex ID). Send id or url. Example: 10.48550/arXiv.1706.03762.
    include_pagesboolean
    Set true to also return markdown split by page. Example: false.
    modelstring
    Model tier for parsing: auto (default) picks one; nano, mini, pro and max trade speed and credits for accuracy. Example: auto.one of autonanominipromaxdefault auto
    urlstring
    Publication page URL from a known provider (DOI, arXiv, PubMed, Semantic Scholar, Springer, Nature, bioRxiv and others). Send id or url. Example: https://arxiv.org/abs/1706.03762.
  • pdfvector_academic_search_grantsSearch research grants and funding opportunities across Grants.gov, NIH RePORTER, CORDIS and UKRI.Read-only

    Search Research Grants

    Search research grants and funding opportunities across Grants.gov, NIH RePORTER, CORDIS and UKRI. Costs 2 credits. Returns grants with title, agency, program, funding amounts, deadlines, eligibility, description and links. Use this to find funding. Use pdfvector_academic_search for publications.

    Inputs

    querystringrequired
    Search query. Example: machine learning for climate modeling.
    deadline_fromstring
    Only grants with a deadline on or after this date (YYYY-MM-DD). Example: 2026-01-01.
    deadline_tostring
    Only grants with a deadline on or before this date (YYYY-MM-DD). Example: 2026-12-31.
    fieldsarray
    Grant fields to return. All fields except providerData are returned when empty. Example: ["title", "agency", "fundingAmountMax", "deadlineDate", "url"]
    funding_maxnumber
    Maximum funding amount. Example: 500000.
    funding_minnumber
    Minimum funding amount. Example: 50000.
    limitinteger
    Results per database, 1 to 50. Defaults to 10. Example: 10.
    offsetinteger
    Results to skip, for paging. Defaults to 0. Example: 0.
    providersarray
    Grant databases to search. Defaults to all four. Example: ["grants-gov", "nih-reporter"]
  • pdfvector_academic_similar_papersFind papers similar to a seed paper using citation network analysis.Read-only

    Find Similar Papers

    Find papers similar to a seed paper using citation network analysis. Costs 3 credits. Returns papers ranked by similarity score with metadata, and citation edges when requested. Use this for related reading. Use pdfvector_academic_paper_graph for direct citations and references.

    Inputs

    idstringrequired
    Seed paper identifier: DOI, Semantic Scholar, arXiv, PubMed or OpenAlex ID, or a URL. Example: 10.48550/arXiv.1706.03762.
    fieldsarray
    Publication fields to return. All fields except providerData are returned when empty. Example: ["title", "doi", "authors", "year", "abstract", "pdfURL"]
    include_edgesboolean
    Set true to include citing and cited-by edges for drawing a graph. Example: false.
    limitinteger
    Maximum similar papers to return, 1 to 100. Defaults to 30. Example: 30.
  • pdfvector_bank_statement_askAnswer a natural-language question about a bank statement with PDF Vector's AI.Read-only

    Ask Bank Statement

    Answer a natural-language question about a bank statement with PDF Vector's AI. Returns the answer text with the model tier used and credits charged. Use this for one question. Use pdfvector_bank_statement_parse for the full text or pdfvector_bank_statement_extract for structured JSON.

    Inputs

    questionstringrequired
    The question to answer about the bank statement. Example: What is the total amount due and when is it due?.
    base64string
    The bank statement file content, base64-encoded, when it has no public URL. Send url or base64. Example: JVBERi0xLjQKJcfsj6IKNSAwIG9iago8PC9MZW5ndGggNiAwIFI+PgpzdHJlYW0K.
    callback_urlstring
    Webhook URL for asynchronous processing. When set, PDF Vector returns 202 at once and POSTs the full result to this URL. Example: https://example.com/hooks/pdfvector.
    document_idstring
    Your own reference ID for this document, used for usage attribution. Example: order-1042.
    modelstring
    Model tier for answering: auto (default) picks one; nano, mini, pro and max trade speed and credits for accuracy. Example: auto.one of autonanominipromaxdefault auto
    urlstring
    Public URL of the bank statement file to process. Send url or base64. Example: https://example.com/statements/2026-10.pdf.