AI Data Extraction
Pricing
from $10.00 / 1,000 document processeds
Extract custom data from PDF, DOCX, and images using AI. Turn invoices, forms, and other documents into structured, validated JSON. Simply provide document URLs and define the fields you need.
You can access the AI Data Extraction programmatically from your own applications by using the Apify API. You can also choose the language preference from below. To use the Apify API, you’ll need an Apify account and your API token, found in API & Integrations in Apify Console.
{ "openapi": "3.0.1", "info": { "version": "0.0", "x-build-id": "uEfZLKkOZ0OWa3sIS" }, "servers": [ { "url": "https://api.apify.com/v2" } ], "paths": { "/acts/obaskly~ai-data-extraction-api/run-sync-get-dataset-items": { "post": { "operationId": "run-sync-get-dataset-items-obaskly-ai-data-extraction-api", "x-openai-isConsequential": false, "summary": "Executes an Actor, waits for its completion, and returns Actor's dataset items in response.", "tags": [ "Run Actor" ], "requestBody": { "required": true, "content": { "application/json": { "schema": { "$ref": "#/components/schemas/inputSchema" } } } }, "parameters": [ { "name": "token", "in": "query", "required": true, "schema": { "type": "string" }, "description": "Enter your Apify token here" } ], "responses": { "200": { "description": "OK" } } } }, "/acts/obaskly~ai-data-extraction-api/runs": { "post": { "operationId": "runs-sync-obaskly-ai-data-extraction-api", "x-openai-isConsequential": false, "summary": "Executes an Actor and returns information about the initiated run in response.", "tags": [ "Run Actor" ], "requestBody": { "required": true, "content": { "application/json": { "schema": { "$ref": "#/components/schemas/inputSchema" } } } }, "parameters": [ { "name": "token", "in": "query", "required": true, "schema": { "type": "string" }, "description": "Enter your Apify token here" } ], "responses": { "200": { "description": "OK", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/runsResponseSchema" } } } } } } }, "/acts/obaskly~ai-data-extraction-api/run-sync": { "post": { "operationId": "run-sync-obaskly-ai-data-extraction-api", "x-openai-isConsequential": false, "summary": "Executes an Actor, waits for completion, and returns the OUTPUT from Key-value store in response.", "tags": [ "Run Actor" ], "requestBody": { "required": true, "content": { "application/json": { "schema": { "$ref": "#/components/schemas/inputSchema" } } } }, "parameters": [ { "name": "token", "in": "query", "required": true, "schema": { "type": "string" }, "description": "Enter your Apify token here" } ], "responses": { "200": { "description": "OK" } } } } }, "components": { "schemas": { "inputSchema": { "type": "object", "required": [ "documents", "fields" ], "properties": { "documents": { "title": "Documents", "type": "array", "description": "Documents to process. Each item needs a publicly accessible HTTPS `url` and may have an optional `filename`. Supported: PDF, DOCX, JPG, JPEG, PNG, WEBP. Free Apify plan: runs are limited to 3 documents and 10 pages in total (documents beyond that are skipped, not charged). You pay only for documents that are successfully extracted; see the Pricing section." }, "fields": { "title": "Fields to extract", "type": "array", "description": "The information you want from every document. Each item has a `name` (letters, digits, underscore), an optional `description` that tells the AI what to look for, and a `type`: string, number, integer, boolean, date, array or object. For array/object fields, describe the structure in the description." }, "language": { "title": "Document language", "type": "string", "description": "Use `auto` to detect the language automatically, or give a code such as `en`, `fr`, `de`, `es`. It improves OCR for scanned documents.", "default": "auto" }, "include_confidence": { "title": "Include confidence estimates", "type": "boolean", "description": "Add a per-field confidence estimate (0 to 1). These are the model's own estimates, not statistically calibrated probabilities.", "default": true }, "include_source_references": { "title": "Include source references", "type": "boolean", "description": "Add the page number and the supporting text for each extracted value. A reference is marked `text_verified` only when the quoted text was found in the document's extracted text.", "default": true }, "vision_mode": { "title": "Visual reading of pages", "enum": [ "auto", "never", "always" ], "type": "string", "description": "`auto`: scanned pages are sent to the AI as an image plus OCR text (most accurate for tables), text pages as text only. `never`: text and OCR only (cheapest, may miss table content in scans). `always`: every page is also sent as an image (highest accuracy and cost).", "default": "auto" }, "max_file_size_mb": { "title": "Maximum file size (MB)", "minimum": 1, "maximum": 50, "type": "integer", "description": "Documents larger than this are rejected without being processed.", "default": 20 }, "max_pages": { "title": "Maximum pages per PDF", "minimum": 1, "maximum": 100, "type": "integer", "description": "PDFs with more pages than this are rejected. Higher limits increase processing time and AI cost.", "default": 30 }, "max_output_tokens": { "title": "Maximum AI output tokens", "minimum": 256, "maximum": 16000, "type": "integer", "description": "Upper bound on the AI reply per document. Leave empty to size it automatically from the number of fields (about 120 tokens per field, up to 16,000). Reasoning tokens count towards this limit. If a document fails with `AI_OUTPUT_TRUNCATED`, raise this or request fewer fields." }, "output_format": { "title": "Output format", "enum": [ "json" ], "type": "string", "description": "Format of the extracted data. Only JSON is currently supported.", "default": "json" } } }, "runsResponseSchema": { "type": "object", "properties": { "data": { "type": "object", "properties": { "id": { "type": "string" }, "actId": { "type": "string" }, "userId": { "type": "string" }, "startedAt": { "type": "string", "format": "date-time", "example": "2025-01-08T00:00:00.000Z" }, "finishedAt": { "type": "string", "format": "date-time", "example": "2025-01-08T00:00:00.000Z" }, "status": { "type": "string", "example": "READY" }, "meta": { "type": "object", "properties": { "origin": { "type": "string", "example": "API" }, "userAgent": { "type": "string" } } }, "stats": { "type": "object", "properties": { "inputBodyLen": { "type": "integer", "example": 2000 }, "rebootCount": { "type": "integer", "example": 0 }, "restartCount": { "type": "integer", "example": 0 }, "resurrectCount": { "type": "integer", "example": 0 }, "computeUnits": { "type": "integer", "example": 0 } } }, "options": { "type": "object", "properties": { "build": { "type": "string", "example": "latest" }, "timeoutSecs": { "type": "integer", "example": 300 }, "memoryMbytes": { "type": "integer", "example": 1024 }, "diskMbytes": { "type": "integer", "example": 2048 } } }, "buildId": { "type": "string" }, "defaultKeyValueStoreId": { "type": "string" }, "defaultDatasetId": { "type": "string" }, "defaultRequestQueueId": { "type": "string" }, "buildNumber": { "type": "string", "example": "1.0.0" }, "containerUrl": { "type": "string" }, "usage": { "type": "object", "properties": { "ACTOR_COMPUTE_UNITS": { "type": "integer", "example": 0 }, "DATASET_READS": { "type": "integer", "example": 0 }, "DATASET_WRITES": { "type": "integer", "example": 0 }, "KEY_VALUE_STORE_READS": { "type": "integer", "example": 0 }, "KEY_VALUE_STORE_WRITES": { "type": "integer", "example": 1 }, "KEY_VALUE_STORE_LISTS": { "type": "integer", "example": 0 }, "REQUEST_QUEUE_READS": { "type": "integer", "example": 0 }, "REQUEST_QUEUE_WRITES": { "type": "integer", "example": 0 }, "DATA_TRANSFER_INTERNAL_GBYTES": { "type": "integer", "example": 0 }, "DATA_TRANSFER_EXTERNAL_GBYTES": { "type": "integer", "example": 0 }, "PROXY_RESIDENTIAL_TRANSFER_GBYTES": { "type": "integer", "example": 0 }, "PROXY_SERPS": { "type": "integer", "example": 0 }, "PROXY_UNBLOCKER_UNITS": { "type": "integer", "example": 0 } } }, "usageTotalUsd": { "type": "number", "example": 0.00005 }, "usageUsd": { "type": "object", "properties": { "ACTOR_COMPUTE_UNITS": { "type": "integer", "example": 0 }, "DATASET_READS": { "type": "integer", "example": 0 }, "DATASET_WRITES": { "type": "integer", "example": 0 }, "KEY_VALUE_STORE_READS": { "type": "integer", "example": 0 }, "KEY_VALUE_STORE_WRITES": { "type": "number", "example": 0.00005 }, "KEY_VALUE_STORE_LISTS": { "type": "integer", "example": 0 }, "REQUEST_QUEUE_READS": { "type": "integer", "example": 0 }, "REQUEST_QUEUE_WRITES": { "type": "integer", "example": 0 }, "DATA_TRANSFER_INTERNAL_GBYTES": { "type": "integer", "example": 0 }, "DATA_TRANSFER_EXTERNAL_GBYTES": { "type": "integer", "example": 0 }, "PROXY_RESIDENTIAL_TRANSFER_GBYTES": { "type": "integer", "example": 0 }, "PROXY_SERPS": { "type": "integer", "example": 0 }, "PROXY_UNBLOCKER_UNITS": { "type": "integer", "example": 0 } } } } } } } } }}OpenAPI is a standard for designing and describing RESTful APIs, allowing developers to define API structure, endpoints, and data formats in a machine-readable way. It simplifies API development, integration, and documentation.
OpenAPI is effective when used with AI agents and GPTs by standardizing how these systems interact with various APIs, for reliable integrations and efficient communication.
By defining machine-readable API specifications, OpenAPI allows AI models like GPTs to understand and use varied data sources, improving accuracy. This accelerates development, reduces errors, and provides context-aware responses, making OpenAPI a core component for AI applications.
You can download the OpenAPI definitions for AI Data Extraction from the options below:
If you’d like to learn more about how OpenAPI powers GPTs, read our blog post.
You can also check out our other API clients: