Extract

POST /extract

Try it

Extract

Python

import requests

url = "https://platform.reducto.ai/extract"

payload = {
    "input": "<string>",
    "parsing": {
        "enhance": {
            "agentic": [],
            "summarize_figures": True
        },
        "retrieval": {
            "chunking": { "chunk_mode": "disabled" },
            "embedding_optimized": False,
            "filter_blocks": []
        },
        "formatting": {
            "add_page_markers": False,
            "include": [],
            "merge_tables": False,
            "table_output_format": "dynamic"
        },
        "spreadsheet": {
            "clustering": "accurate",
            "exclude": [],
            "include": [],
            "split_large_tables": {
                "enabled": True,
                "size": 50
            }
        },
        "settings": {
            "embed_pdf_metadata": False,
            "force_url_result": False,
            "ocr_system": "standard",
            "persist_results": False,
            "return_images": [],
            "return_ocr_data": False
        }
    },
    "instructions": {
        "schema": {},
        "system_prompt": "Be precise and thorough."
    },
    "settings": {
        "include_images": False,
        "optimize_for_latency": False,
        "array_extract": False,
        "citations": {
            "enabled": False,
            "numerical_confidence": True
        }
    }
}
headers = {
    "Authorization": "Bearer <token>",
    "Content-Type": "application/json"
}

response = requests.post(url, json=payload, headers=headers)

print(response.json())

200

{
  "usage": {
    "num_pages": 123,
    "num_fields": 123,
    "credits": 123
  },
  "result": "<unknown>",
  "job_id": "<string>",
  "studio_link": "<string>"
}

Authorizations

Authorization · string · header · required

Bearer authentication header of the form Bearer <token>, where <token> is your auth token.

Body

application/json

SyncExtractConfig

input · required

For parse/split/extract pipelines, the URL of the document to be processed. You can provide one of the following: 1. A publicly available URL 2. A presigned S3 URL 3. A reducto:// prefixed URL obtained from the /upload endpoint after directly uploading a document 4. A jobid:// prefixed URL obtained from a previous /parse invocation

For edit pipelines, this should be a string containing the edit instructions

parsing · ParseOptions · object

The configuration options for parsing the document. If you are passing in a jobid:// URL for the file, then this configuration will be ignored.

Show child attributes

parsing. enhance · Enhance · object

Show child attributes

parsing.enhance. agentic · (TableAgentic · object | FigureAgentic · object | TextAgentic · object)[]

Agentic uses vision language models to enhance the accuracy of the output of different types of extraction. This will incur a cost and latency increase.

TableAgentic

Show child attributes

parsing.enhance.agentic. scope · string · required

Allowed value: "table"

parsing.enhance.agentic. prompt · string | null

Custom prompt for table agentic.

parsing.enhance. summarize_figures · boolean · default: true

If True, summarize figures using a small vision language model. Defaults to True.

parsing. retrieval · Retrieval · object

Show child attributes

parsing.retrieval. chunking · Chunking · object

Show child attributes

parsing.retrieval.chunking. chunk_mode · enum<string> · default: disabled

Choose how to partition chunks. Variable mode chunks by character length and visual context. Section mode chunks by section headers. Page mode chunks according to pages. Page sections mode chunks first by page, then by sections within each page. Disabled returns one single chunk.

Available options: variable, section, page, disabled, block, page_sections

parsing.retrieval.chunking. chunk_size · integer | null

The approximate size of chunks (in characters) that the document will be split into. Defaults to null, in which case the chunk size is variable between 250 - 1500 characters.

parsing.retrieval. filter_blocks · enum<string>[]

A list of block types to filter out from 'content' and 'embed' fields. By default, no blocks are filtered.

Available options: Header, Footer, Title, Section Header, Page Number, List Item, Figure, Table, Key Value, Text, Comment, Signature

parsing.retrieval. embedding_optimized · boolean · default: false

If True, use embedding optimized mode. Defaults to False.

parsing. formatting · Formatting · object

Show child attributes

parsing.formatting. add_page_markers · boolean · default: false

If True, add page markers to the output. Defaults to False. Useful for extracting data with page specific information.

parsing.formatting. table_output_format · enum<string> · default: dynamic

The mode to use for table output. Defaults to dynamic, which returns md for simpler tables and html for more complex tables.

Available options: html, json, md, jsonbbox, dynamic, csv

parsing.formatting. merge_tables · boolean · default: false

A flag to indicate if consecutive tables with the same number of columns should be merged. Defaults to False.

parsing.formatting. include · enum<string>[]

A list of formatting to include in the output. [insert description of each option here later]

Available options: change_tracking, highlight, comments, hyperlinks

parsing. spreadsheet · Spreadsheet · object

Show child attributes

parsing.spreadsheet. split_large_tables · SplitLargeTables · object

Show child attributes

parsing.spreadsheet.split_large_tables. enabled · boolean · default: true

If True, split large tables into smaller tables. Defaults to True.

parsing.spreadsheet.split_large_tables. size · integer · default: 50

The size of the tables to split into. Defaults to 50.

parsing.spreadsheet. include · enum<string>[]

Whether to include cell color and formula information in the output.

Available options: cell_colors, formula

parsing.spreadsheet. clustering · enum<string> · default: accurate

In a spreadsheet with different tables inside, we enable splitting up the tables by default. Accurate mode applies more powerful models for superior accuracy, at 5× the default per-cell rate. Disabling will register as one large table.

Available options: accurate, fast, disabled

parsing.spreadsheet. exclude · enum<string>[]

Whether to exclude hidden sheets, rows, or columns in the output.

Available options: hidden_sheets, hidden_rows, hidden_cols

parsing. settings · Settings · object

Show child attributes

parsing.settings. ocr_system · enum<string> · default: standard

Standard is our best multilingual OCR system. Legacy only supports germanic languages and is available for backwards compatibility.

Available options: standard, legacy

parsing.settings. force_url_result · boolean · default: false

Force the result to be returned in URL form.

parsing.settings. force_file_extension · string | null

Force the URL to be downloaded as a specific file extension (e.g. .png).

parsing.settings. return_ocr_data · boolean · default: false

If True, return OCR data in the result. Defaults to False.

parsing.settings. return_images · enum<string>[]

Whether to return images for the specified block types. By default, no images are returned.

Available options: figure, table

parsing.settings. embed_pdf_metadata · boolean · default: false

If True, embed OCR metadata into the returned PDF. Defaults to False.

parsing.settings. persist_results · boolean · default: false

If True, persist the results indefinitely. Defaults to False.

parsing.settings. timeout · number | null

The timeout for the job in seconds.

parsing.settings. page_range

The page range to process (1-indexed). By default, the entire document is processed.

Show child attributes

parsing.settings.page_range. start · integer | null

The page number to start processing from (1-indexed).

parsing.settings.page_range. end · integer | null

The page number to stop processing at (1-indexed).

parsing.settings. document_password · string | null

Password to decrypt password-protected documents.

instructions · Instructions · object

The instructions to use for the extraction.

Show child attributes

instructions. schema · any · default: {}

The JSON schema to use for the extraction.

instructions. system_prompt · string · default: Be precise and thorough.

The system prompt to use for the extraction.

settings · ExtractSettings · object

The settings to use for the extraction.

Show child attributes

settings. include_images · boolean · default: false

If True, include images in the extraction.

settings. optimize_for_latency · boolean · default: false

If True, jobs will be processed with a higher throughput and priority at a higher cost. Defaults to False.

settings. array_extract · boolean · default: false

If True, use array extraction.

settings. citations · Citations · object

The citations to use for the extraction.

Show child attributes

settings.citations. enabled · boolean · default: false

If True, include citations in the extraction.

settings.citations. numerical_confidence · boolean · default: true

If True, enable numeric citation confidence scores. Defaults to True.

Response

200

application/json

Successful Response

V3ExtractResponse

usage · ExtractUsage · object · required

Show child attributes

usage. num_pages · integer · required

usage. num_fields · integer · required

usage. credits · number | null

result · required

The extracted response in your provided schema. This is a list of dictionaries. If disable_chunking is True (default), then it will be a list of length one.

job_id · string | null

studio_link · string | null

The link to the studio pipeline for the document.