List extractions for a resolution result

The individual extractions that contributed to a resolution result, including the raw extracted value, the verbatim source passage, the AI’s reasoning, and consistency metadata. For document extractions, document_id and page_number indicate where the value was pulled from; for web extractions, web_search_query and web_citations describe the source.

Results are ordered with web sources first, then consistent rows, then by document and page number, for stable pagination as new extractions arrive.

On this page

HTTP request 

GET https://api.parsewise.ai/api/v1/projects/{project_id}/results/{resolution_result_id}/extractions/

URI Parameters

Name In Required Type Description
limit query No integer Maximum number of extractions to return on this page. Defaults to 10.
offset query No integer Zero-indexed offset into the ordered list of extractions, used for pagination. Defaults to 0.
project_id path Yes string (uuid)  
resolution_result_id path Yes string (uuid)  

Request Header

Name Required Type Description
X-API-Key Yes string API key with the pw_live_ prefix. See Authentication.

Responses

Status Type Description
200 V1ExtractionsListResponse —

Security

  • ApiKeyAuth — apiKey — in X-API-Key header. API key with pw_live_ prefix.

Python example 

import os
import requests

API_KEY = os.environ["PARSEWISE_API_KEY"]
BASE_URL = "https://api.parsewise.ai/api/v1"

project_id = "<uuid>"
resolution_result_id = "<uuid>"

params = {
    # "limit": 0,  # Maximum number of extractions to return on this page
    # "offset": 0,  # Zero-indexed offset into the ordered list of extractions, used for pagination
}

resp = requests.get(
    f"{BASE_URL}/projects/{project_id}/results/{resolution_result_id}/extractions/",
    headers={"X-API-Key": API_KEY},
    params=params,
)
resp.raise_for_status()
print(resp.json() if resp.content else None)

Definitions

V1ExtractionResult

A single extraction underlying a resolution result.

One resolution result is typically produced from many extractions across different document pages (and optionally web sources). This is one such extraction, with the raw value the AI pulled from the source plus the surrounding text and consistency metadata.

Name Required Type Description
id Yes string (uuid) Stable identifier of the extraction. Use this with GET /projects/{project_id}/extractions/{extraction_id}/bounding-boxes/ to retrieve where on the page the value was found.
value Yes string Raw extracted value as a string (numeric values are normalized to a canonical decimal representation). For the resolved, typed value across all extractions for a result, see the parent resolution result.
relevant_passage Yes string Verbatim passage from the source page that the AI used to derive the extracted value.
reasoning Yes string AI-generated explanation of how the value was derived from the source passage.
consistent Yes boolean (nullable) Whether this extraction agrees with the resolved value. null when consistency has not yet been evaluated.
justification_for_consistency Yes string (nullable) AI-generated explanation of the consistency outcome. null when consistency has not yet been evaluated.
source_type Yes V1ExtractionResultSourceTypeEnum Whether the extraction came from an uploaded document or a web search result.
document_id Yes string (uuid) (nullable) Identifier of the source document. null for web extractions.
document_name Yes string (nullable) Stored filename of the source document. null for web extractions.
page_number Yes integer (nullable) 1-indexed page number within the source document. null for web extractions.
extracted_at Yes string (date-time) When the extraction was first produced.
web_search_query Yes string (nullable) Web search query that surfaced this extraction. Populated only for web extractions; null for document extractions.
web_search_timestamp Yes string (date-time) (nullable) When the web search was executed. Populated only for web extractions; null for document extractions.
web_citations Yes array<WebCitation> Web pages cited by this extraction along with the snippets of supporting text. Empty list for document extractions.

V1ExtractionResultSourceTypeEnum

Type: string. One of:

  • document
  • web

V1ExtractionsListResponse

Name Required Type Description
data Yes array<V1ExtractionResult> Extractions on this page, ordered with web sources first, then consistent rows, then by document and page.
total_count Yes integer Total number of extractions for this resolution result across all pages.

WebCitation

One web citation: the snippets of supporting text and the source page.

Name Required Type Description
snippets Yes array<WebSnippet> Snippets of text from the source page that support the extracted value.
source Yes WebSource The web page these snippets came from.

WebSnippet

One snippet of supporting text quoted from a web source page.

Name Required Type Description
text Yes string Verbatim snippet of text from the source page.

WebSource

A single web page (URL + title) cited by an extraction.

Name Required Type Description
url Yes string (uri) URL of the cited web page.
title Yes string Title of the cited web page.