Entity intelligence resolves a company, person, product or creator to one canonical record and measures what is known, said and believed about it across ChatGPT (OpenAI), Claude (Anthropic), Gemini (Google DeepMind), Perplexity (Perplexity AI), Grok (xAI) and the open web — Observations carrying source, timestamp and confidence on every fact, not a raw text dump.
Every record is verified, signed, and correctable — over REST, a hosted MCP server, and a TypeScript SDK.
Nine data products over one resolved entity record.
Resolve a name, URL or partial match to a stable canonical ID with a confidence band.
One call for identity, categorization, headline signals, coverage, scores and relationships.
The three headline signals — how often, how favourably, and what share of the conversation.
How much of what each AI engine says about the entity matches the verified record, and what it gets wrong.
How long a correction takes to reach each engine, measured from publication to adoption.
Named, timestamped shifts in the entity picture — breakout, erosion, authority lag, crisis.
Cohort rank and score deltas versus the cohort median, side by side.
A creator or podcast name, or any one profile link, resolved to every cross-verified handle.
Legal-entity registry data and, where public, listing and quote data for the entity or its owner.
The same resolved entity record is available wherever your stack needs to read it — a direct REST call, an agent over MCP, or a typed SDK client.
The full data product set under /api/v1 — JSON over HTTPS, keyed by account, with deeper layers unlocking by plan.
A hosted Model Context Protocol endpoint — agents query and correct verified entity data directly, no self-hosting required.
Connect the MCP server →@entidex/sdk on npm — typed methods mirroring every REST endpoint, ready to drop into a Node or edge runtime.
A native command-line client is in private beta. Every CLI command documented has a working REST and MCP equivalent today.
One call turns a raw name into a canonical, confidence-scored entity ID.
curl -s "https://entidex.com/api/v1/entities/resolve?q=anthropic" \
-H "Authorization: Bearer entx_live_sk_..."Most tools stop at telling you how often AI mentions an entity. Entidex also tells you how much of what each AI engine says is wrong — a per-engine accuracy read against a signed, verified record, with the exact fields each engine gets wrong or is simply unaware of. When a correction is published, Information Lag measures how long it takes each engine to adopt it. Detect the gap, publish the verified record, measure adoption — the full correction loop, not a snapshot.
Explore the dataset catalog →An entity intelligence API resolves a real-world entity — a company, person, product or creator — to one canonical, provenance-backed record, and returns what is known, said and believed about it across AI engines and the open web. Instead of raw text or a search result, you get a stable ID plus structured, sourced signals you can query and compare over time.
A data enrichment API appends firmographic fields to a record you already have — headcount, industry code, domain. An entity intelligence API returns Observations, each carrying its source surface, timestamp and confidence, plus derived signals — AI Visibility, Sentiment, Share of Voice and per-engine accuracy — and a verified record you can correct when an AI engine gets it wrong.
Entidex tracks the five major AI engines: ChatGPT (OpenAI), Claude (Anthropic), Gemini (Google DeepMind), Perplexity (Perplexity AI), Grok (xAI). Every entity record includes per-engine visibility, sentiment and accuracy so you can see where each engine agrees, diverges or is simply wrong.
Yes. A free API key gets rate-limited entity resolution and entity lookup with cached headline signals. The free scan and Creator Resolution are keyless — no account required to see what AI currently says about an entity.
Free API key, no card required. Or see the full tier map first.