Google Knowledge Graph

Turn any list of entity names — companies, people, places, products — into structured metadata pulled straight from the Google Knowledge Graph. No scraping, no parsing, no guessing.

What you get

• One dataset row per match with entity ID, canonical name, schema.org types, short and long descriptions, official URL, and image URLs
• Configurable match limit (1-20) per query — top match by default, or fetch the full candidate list for disambiguation
• Type filtering to constrain results to Person, Organization, Place, or any schema.org type
• Multi-language support — request results in any BCP-47 language
• Result scores for confidence-weighted downstream logic
• Error-tolerant — failed lookups produce an error row and the run continues

Example output

{
  "query": "Anthropic",
  "rank": 1,
  "status": "success",
  "entityId": "kg:/g/11h3pnwf3p",
  "name": "Anthropic",
  "types": ["Corporation", "Organization", "Thing"],
  "description": "Artificial intelligence company",
  "detailedDescription": "Anthropic is an American artificial intelligence startup company...",
  "detailedDescriptionUrl": "https://en.wikipedia.org/wiki/Anthropic",
  "imageUrl": "https://upload.wikimedia.org/...png",
  "imageContentUrl": "https://en.wikipedia.org/wiki/File:Anthropic_logo.svg",
  "url": "https://www.anthropic.com/",
  "resultScore": 1432.51,
  "error": null
}

Use cases

• CRM enrichment — turn a column of company names into a structured profile with logos, descriptions, and official URLs
• Lead scoring — use resultScore and entity types to filter out generic terms and prioritise real organisations
• "Is this a real company?" verification — separate well-known entities from misspellings, shell companies, or unknown brands before outreach
• Content categorisation — tag articles, products, or tickets with schema.org types from authoritative entity matches
• Entity disambiguation for ML pipelines — resolve ambiguous strings ("Apple", "Mercury") to a canonical Knowledge Graph ID before downstream processing
• Knowledge base seeding — bootstrap a graph of people, companies, and places with descriptions and images sourced from Wikipedia and Wikidata

How to use

1. Paste your entity names into the Queries field — one per entry
2. Set Limit (default 1 — top match only; raise to 5-10 for disambiguation workflows)
3. (Optional) Restrict Types to filter results, e.g. Organization, Person, Place
4. (Optional) Set Languages for localised names and descriptions (default en)
5. Run the actor — every match appears as its own dataset row, ready to export as CSV or JSON

Output schema

Field	Type	Description
query	string	The input query
rank	number	1-based rank within the query (0 if no matches or error)
status	string	"success" or "error"
entityId	string | null	Knowledge Graph ID (e.g. kg:/m/0d3kx)
name	string | null	Canonical entity name
types	string[]	Schema.org @type values
description	string | null	Short description
detailedDescription	string | null	Long-form description (typically Wikipedia intro)
detailedDescriptionUrl	string | null	Source URL of the detailed description
imageUrl	string | null	Direct image URL (image.contentUrl)
imageContentUrl	string | null	Image source page (image.url)
url	string | null	Entity's official URL
resultScore	number | null	Google's relevance score
error	string | null	Error message if status is "error", or "no matches"

If a query fails, status is "error" and error contains the reason — the run continues with the remaining queries.

Tips

• For CRM enrichment, set Limit to 1 — you almost always want the top match
• For disambiguation, set Limit to 5-10 and inspect resultScore and types
• Filter by types: ["Organization"] to skip people and places when enriching company lists
• The Google Knowledge Graph favours well-known entities — obscure local businesses may return no matches