Google Knowledge Graph Search API

The Google Knowledge Graph Search API lets you look up entities in Google’s Knowledge Graph programmatically. You send a name or ID and get back matching entities with their name, schema.org types, a short description, a longer description (often from Wikipedia), the entity’s Knowledge Graph ID and a relevance score. For SEOs, it is the quickest way to check whether Google recognises an entity and how it describes it.

How it works

You call the entities:search endpoint with an API key and a query, optionally filtering by type (such as Organization or Person), language or a known ID. The response is JSON-LD that follows schema.org. Each result includes a resultScore, which indicates how well that entity matches your request.

Two warnings from Google’s documentation matter. The API is being migrated to Cloud Enterprise Knowledge Graph, which Google recommends for new projects, and it is not suitable as a production-critical service.

How it’s used in SEO and marketing

• Entity recognition checks. Query your brand and key people. If nothing comes back, or a different entity with the same name does, you have a disambiguation problem.
• Competitor diagnosis. See which entities and descriptions Google associates with competitors compared with you.
• Keyword expansion. Pull related entities for a topic and use them to widen a keyword list.
• Keyword grouping. Cluster keywords by the entities they share, and visualise the relationships as a network graph.
• Schema checks. Confirm that the entity your markup describes matches what Google already holds.

Example

MLforSEO’s Knowledge Graph notebook takes a list of seed terms, queries the API for each, and stores the name, type, description and ID of every match. Grouping the results by entity type shows which seeds map to well-established entities and which return nothing, a direct pointer to topics where Google’s understanding is thin and clear, well-structured content can help.

Limitations and things to watch

• The API returns a subset of Knowledge Graph data and does not reproduce a knowledge panel.
• resultScore is a match signal for that request. It is not search volume or popularity, and scores are not comparable across different queries.
• Being present in the Knowledge Graph does not guarantee a knowledge panel, and AI assistants draw on other sources too.
• Plan for the migration notice: keep your code easy to switch to Cloud Enterprise Knowledge Graph.

Learn it on MLforSEO

• What Is a Knowledge Graph? How to Use It for Smarter Keyword Strategy
• Knowledge Graph API Extraction and Entity Relationship Analysis (Notebook)
• 5 Steps to Identify and Fix Entity Recognition Gaps
• Semantic AI-powered SEO Keyword Research course
• AI Search and LLMs: Entity SEO and Knowledge Graph Strategies for Brands course