Mindee

Mindee is a document parsing and intelligent document processing platform. Its REST API (api-v2.mindee.net) provides asynchronous inference across five model families — Extraction, Classification, Crop, OCR, and Split — plus a unified Jobs endpoint for polling. Mindee ships prebuilt models for invoices, receipts, passports, IDs, resumes, bank statements, and barcodes, along with customer-configurable extraction models with their own Data Schemas. Native SDKs cover Python, Node.js, Java, PHP, Ruby, and .NET, and front-end vision SDKs cover React and Vue. Mindee also stewards the open-source docTR OCR library.

6 APIs 15 Features

Document ParsingOCRIDPAIMachine LearningInvoicesReceiptsIDsComputer Vision

APIs

Mindee Extraction API

Extract structured fields from invoices, receipts, IDs, passports, resumes, bank statements, and any other document type via Mindee's configurable extraction models. Enqueue a f...

Mindee Classification API

Classify documents into predefined categories so they can be routed to the appropriate downstream extraction model. Enqueue at POST /v2/products/classification/enqueue and retri...

Mindee Crop API

Detect, isolate, and crop the document area inside a larger scanned image. Useful for cleaning up phone-captured documents prior to classification or extraction. Async enqueue/r...

Mindee OCR API

Run full-page optical character recognition on a document and return the raw text plus bounding polygons. Enqueue at POST /v2/products/ocr/enqueue and retrieve the OCR payload f...

Mindee Split API

Split a multi-document PDF or image batch into individual documents that can each be classified and extracted independently. Async enqueue/results pattern at /v2/products/split/.

Mindee Jobs API

Poll the status of asynchronous Mindee inference jobs across every Mindee model family via GET /v2/jobs/{job_id}. Jobs return Waiting, Processing, Failed, or Success states and ...

Collections

Mindee Classification API

POSTMAN

Mindee Crop API

POSTMAN

Mindee Extraction API

POSTMAN

Mindee Jobs API

POSTMAN

Mindee OCR API

POSTMAN

Mindee Split API

POSTMAN

Mindee Classification API

OPEN

Mindee Crop API

OPEN

Mindee Extraction API

OPEN

Mindee Jobs API

OPEN

Mindee OCR API

OPEN

Mindee Split API

OPEN

Arazzo Workflows

Mindee Bank Statement Extraction

Enqueue a bank statement with RAG and raw text enabled, poll until processed, then read the extracted transactions and balances.

ARAZZO

Mindee Classify Document Type

Enqueue a document for classification, poll until processed, then read the predicted document type.

ARAZZO

Mindee Classify Then Extract

Classify an unknown document, then run extraction on the same file once the type is known, reading the parsed fields.

ARAZZO

Mindee Crop Document

Enqueue a document for cropping, poll until processed, then read the detected objects and their crop coordinates.

ARAZZO

Mindee Custom Document Extraction

Enqueue an arbitrary document against a custom model with raw text capture, poll until processed, then read fields and full text.

ARAZZO

Mindee Extract From URL

Enqueue a publicly hosted document by URL for extraction, poll until processed, then read the extracted fields.

ARAZZO

Mindee Invoice Extraction

Enqueue an invoice for extraction, poll the job until processed, then read the extracted fields.

ARAZZO

Mindee OCR Full Text

Enqueue a document for full-page OCR, poll until processed, then read the per-page words and text content.

ARAZZO

Mindee OCR Then Extract

Run OCR over a document to capture its raw text, then extract structured fields from the same file, reading both outputs.

ARAZZO

Mindee Passport and ID Extraction

Enqueue a passport or identity document with polygon locations, poll until processed, then read the extracted holder fields.

ARAZZO

Mindee Poll Job With Failure Branch

Enqueue an extraction, poll the job, and branch explicitly on Processed, Processing, or Failed before reading the result.

ARAZZO

Mindee Receipt Extraction

Enqueue a receipt with confidence scoring, poll until processed, then read the extracted line items and totals.

ARAZZO

Mindee Split Multi-Document File

Enqueue a multi-document file for splitting, poll until processed, then read the identified document ranges.

ARAZZO

Mindee Split Then Extract

Split a multi-document file into ranges, then extract structured fields from the original file, reading splits and fields.

ARAZZO

Pricing Plans

Mindee Plans Pricing

4 plans

PLANS

Rate Limits

Mindee Rate Limits

0 limits

RATE LIMITS

FinOps

Mindee Finops

FINOPS

Features

Async enqueue/results pattern across all Mindee models with a unified job polling endpoint

Configurable extraction models with customer-defined Data Schemas

Library of ready-to-use models for invoices, receipts, IDs, passports, driver's licenses, resumes, bank statements, and barcodes

Classification, Crop, OCR, and Split utility models for full IDP pipelines

Continuous Learning via Retrieval-Augmented Generation against a customer knowledge base

Confidence scores and bounding polygons (Pro and higher)

Full raw OCR text option on extraction inferences

100 MB file size limit; up to 200 pages per PDF

Webhook-based result delivery in addition to polling

Per-model and per-API-key usage insights inside the platform console

Native SDKs for Python, Node.js, Java, PHP, Ruby, and .NET

Front-end computer vision SDKs for React and Vue

Open-source docTR library for self-hosted OCR

Zapier and Microsoft Flow integrations for no-code automation

Credit-based pricing with monthly subscription tiers and per-credit overage

Semantic Vocabularies

Mindee Context

11 classes · 14 properties

JSON-LD

API Governance Rules

Mindee API Rules

6 rules · 3 errors 2 warnings 1 info

SPECTRAL

JSON Structure

Mindee Extraction Structure

1 properties

JSON STRUCTURE

Example Payloads

Mindee Extraction Enqueue Example

2 fields

EXAMPLE

Mindee Extraction Result Example

2 fields

EXAMPLE

Resources

PostmanWorkspace

PostmanWorkspace

Documentation

GettingStarted

Documentation

Authentication

Documentation

Documentation

Authentication

GitHubOrganization

GitHubOrganization

Documentation

SpectralRules

Sources

opencollection: 1.0.0
info:
  name: Mindee Split API
  version: 2.0.0
request:
  auth:
    type: apikey
    key: Authorization
    value: '{{Authorization}}'
    placement: header
items:
- info:
    name: Split
    type: folder
  items:
  - info:
      name: Send A File For A Split Utility Inference
      type: http
    http:
      method: POST
      url: https://api-v2.mindee.net/v2/products/split/enqueue
      body:
        type: json
        data: '{}'
      auth:
        type: apikey
        key: Authorization
        value: '{{Authorization}}'
        placement: header
    docs: "Send a file to the asynchronous processing queue for a split utility inference.\n\nResults can be retrieved in\
      \ two ways:\n* Loop on the `polling_url` endpoint in the response until the status\n  is `Processed` or `Failed` (polling).\n\
      * Use a webhook to be notified when the inference is finished.\n  Webhooks can be configured on the platform on a per-model\
      \ basis.\n\nInferences taking longer than 590 seconds will time out."
  - info:
      name: Get The Result Of A Completed Split Product Inference
      type: http
    http:
      method: GET
      url: https://api-v2.mindee.net/v2/products/split/results/:inference_id
      params:
      - name: inference_id
        value: ''
        type: path
        description: UUID of the inference to retrieve
      auth:
        type: apikey
        key: Authorization
        value: '{{Authorization}}'
        placement: header
    docs: 'Get the result of a split utility inference that was previously enqueued.


      The inference will only be available after it has finished processing.


      If you are using webhooks there is no need to call this route.<br/>

      However, the JSON payload described here is what will be sent to each specified webhook''s URL.


      If you activated the "Delete extracted data when fetched" option for the model,

      the inference will be deleted after this route completes.<br>

      Any further requests to the same inference ID w'
bundled: true