Docparser.dev

Documentation search

Find an answer

Start typing to search the documentation.
    Open portal

    01 PRODUCT SYSTEM

    A document job with a fixed destination.

    Docparser accepts a source URL, selects the document contract before reading it, and returns structured data your system can inspect—not a free-form answer.

    Read the quickstart Open the portal

    DOCUMENT JOB / CONTRACT

    1. 01source URL→
    2. 02typeId + schema→
    3. 03structured payload→
    4. 04webhook or retrieve

    02 PIPELINE

    The schema is chosen before extraction.

    Select a live typeId for the document you are sending. That schema defines the expected shape, including summary fields and line items where the document type supports them.

    1. URL → job

      Submit

      Send a PDF or image URL with your document type.

    2. typeId → schema

      Select schema

      A fixed field contract frames extraction for that document.

    3. source → JSON

      Extract

      The document is read into summary fields and line items.

    4. job → system

      Deliver

      Receive the result at your webhook or retrieve the job.

    03 SUBMISSION MODES

    Send the source in the form the job expects.

    The API infers the mode from the request body: attachments first, then multiple document URLs, then the single-document shape.

    1. 01
      documentUrl

      One document URL

      Submit a PDF or image URL when one source contains one document to process.

    2. 02
      documentUrls[]

      Many image URLs

      Submit ordered image URLs and choose a single-invoice or multiple-invoice processing mode.

    3. 03
      attachments[]

      Email attachments

      Send attachment URLs with the email subject and body for the email-specific workflow.

    04 MEASURED JOB

    Time, shown as a trace—not a promise.

    This early one-job production trace measured 18.4 s end to end. The model call dominated that run; Different documents, sources and upstream conditions will vary.

    Docparser Labs · n = 1 production job · measured 2026-07-24

    One early production-job trace 18.4 s total

    Docparser Labs · n = 1 production job · measured 2026-07-24

    Queue: 15 ms Source download: 525 ms Model call: 14.5 s Webhook: 203 ms ≈3.2 s uninstrumented remainder (derived from rounded total): ≈3.2 s
    • Queue 15 ms
    • Source download 525 ms
    • Model call 14.5 s
    • Webhook 203 ms
    • ≈3.2 s uninstrumented remainder (derived from rounded total) ≈3.2 s

    05 JOB STATES

    A status you can follow.

    Submit returns a job identifier and your consumerRefId. Retrieve by that reference from the organization that created it.

    1. 01
      initiated

      Accepted and queued with your correlation reference.

    2. 02
      extracting

      The selected document contract guides model extraction.

    3. 03
      processed

      Structured data is persisted and ready for delivery or retrieval.

    4. 04
      error

      The job records a failure state and its delivery attempt.

    06 DELIVERY

    Choose push, pull, or both.

    A webhook path is joined to the base URL registered on the API key. Or retrieve the persisted job later with the same organization credentials.

    WEBHOOK Push on completion
    Receive the processed or error payload at the callback path attached to the job.
    RETRIEVE Poll by reference
    Poll by your consumerRefId when your workflow wants to own the timing.