Diffbot crawls the web and structures it into a massive knowledge graph of billions of entities, with extraction APIs that turn pages into normalized data. Entity Enricher takes a different starting point: instead of querying a prebuilt graph of the web as it exists, it enriches the entities you define against your schema, using multiple LLMs plus your documents and live web search — with arbitration and identity built in.
You define the output structure — any fields, any nesting, any entity type — and the models fill it.
You query Diffbot’s fixed knowledge-graph entity types and fields. The shape is theirs.
Your input keys are preserved verbatim and each entity gets a stable, tenant-scoped semantic ID.
Entities are matched to global knowledge-graph IDs, not modeled around your own records.
Two or more LLMs answer in parallel; conflicts are resolved field by field with the reasoning stored.
A single extraction/graph pipeline. No cross-model validation or per-field audit trail.
Best when the entity and schema are yours, and accuracy/identity matter more than raw web coverage.
Best when you need a huge prebuilt graph or to structure arbitrary public web pages at scale.
| Feature | Entity Enricher | Diffbot |
|---|---|---|
| Custom output schema (any fields) | Fixed entity types | |
| Any entity type you define | KG entity types only | |
| Source of truth | LLMs + your docs + web | Web crawl + knowledge graph |
| Enrich your own records / keys | Match to KG entities | |
| Multi-model cross-validation | ||
| Field-level arbitration + audit trail | ||
| Semantic IDs (your-tenant identity) | Global KG IDs | |
| Document ingestion (PDF, DOCX, images) | ||
| Web page / HTML extraction | Via web search | |
| Multilingual output (40 languages) | ||
| Bring your own keys / self-hosted models | ||
| Massive prebuilt entity graph | ||
| REST API + MCP + n8n / Make surfaces | API | |
| Pricing Model | Pay-per-token (BYOK) | Subscription / enterprise |
Pay-per-token
Bring your own LLM API keys and pay your provider directly for tokens consumed. No platform subscription, no per-call graph-query fee.
Subscription
Monthly plans metered by API credits / queries against the knowledge graph and extraction APIs, with enterprise tiers for high-volume crawling and KG access.
Pricing reflects publicly published tiers and can change — check each vendor for current rates.
Skip the fixed graph schema. Define your own structure, enrich any entity from multiple models plus your documents and the web, and keep your own keys and identity.
Get Started Free