Everything you need to build on Entidex: REST API, TypeScript SDK, CLI, reconciliation, decision-survival queries, machine-readable knowledge statements, and the MCP server.
Sign up for a free API key and start reading without a paid plan. A free key can resolve any entity and read the lite Entity Profile — canonical identity, identifiers, and last-cached headline signals (AI Visibility, Sentiment, Share of Voice) — on GET /api/v1/entities/resolve and GET /api/v1/entities/{entityId}. It can also pull the complete, categorized record on GET /api/v1/entities/{entityId}/complete, browse the network on GET /api/v1/categories, and sample raw observations on GET /api/v1/discovery/entity/{entityId}/observations. Rate-limited and read-only.
Free tier key: 100 monthly credits for keyed evaluation reads (per-minute rate limit protects fair use). Developer tier: 1,000/mo (1,500 with verified email + phone), then PAYG at £0.0008/credit. Per-call cost: reads (resolve, complete, grounding-packet, raw observations) = 1 credit · refresh = 100 · decision-survival/run = 500 · full-check = 1,000. Upgrade to Starter for unlimited lite-profile reads, or Developer API (£29/mo) for predictable flat-rate access to L1–L4.
Try it without a key:
curl -X POST https://entidex.com/api/scan \
-H 'Content-Type: application/json' \
-d '{"query":"openai.com"}' | jq .Full profile, claims, insights, drift and alerts are available on paid tiers. See pricing for what each tier unlocks.
Keyless instant value: POST a domain to /api/scan and get the claim ledger — what AI engines say vs the verified record, plus a gap score. No signup, no key.
Your first call, on every workflow: one small, deterministic, signed payload of verified entity facts + corrections to what AI engines get wrong. GET /api/v1/entities/{id}/grounding-packet on a free key; ?format=grounding-metadata for Vertex-shaped pipelines; entity_grounding_packet over MCP; .groundingPacket() in the SDK. Grounding, not enrichment.
OpenAPI spec, tier mapping, endpoint catalog and quick start. The canonical reference.
Official zero-dependency client — npm install @entidex/sdk. resolve, getEntity, complete (the full categorized record in one call), and insights, fully typed against the v1 contract.
Preview the planned UX. The CLI is in private beta — use the REST API or MCP server today.
Resolve messy entity strings (domains, names, aliases) to canonical Entidex IDs in bulk. Confidence-scored, deduped against the live Entidex entity graph. Patterns for batch matching with thresholds, candidate disambiguation, and Wikidata cross-reference.
Resolve any creator name or profile link to the canonical creator — every cross-verified handle, with provenance and a signed record. Keyless demo on the landing page; key-gated dataset twin at /api/v1/creators/resolve.
Query the AI engines under decision pressure and measure whether your entity survives. Endpoint + interpretation guide.
Ground your agents on resolved, provenance-backed entity records over MCP and the API: resolve, verify, read, correct. One layer, any agent.
Guardrail, retriever, and reader recipes for OpenAI agents, LangChain, and LlamaIndex — the verified record inside pipelines you already run. Every sample executed against production before it shipped.
Drop Entidex into any MCP-compatible LLM client. Live entity intelligence inside your AI workflow.
Crawlable, machine-readable verified records (Markdown · schema.org ClaimReview · JSON), signed by did:web:entidex.com, discoverable via llms.txt, verifiable in-flight via the entity_verify_fact MCP tool.