Entity Enricher

What Entity Enricher MCP server does

The Entity Enricher MCP server provides an MCP interface to Entity Enricher at https://entityenricher.ai/api/mcp/. It helps an agent turn sample data, documents, or a direct description into reusable schemas, then apply those schemas to individual entities or lists. Schemas can contain multilingual fields and can be edited at the property level.

The Entity Enricher MCP server also supports combining records, retrying only failed expertise domains, and resolving recurring objects to semantic identities. Identity workflows include managing aliases and reviewing uncertain matches. For teams that need to move structured results into an application, the service can derive relational tables and migrations for PostgreSQL, MySQL, or SQLite.

How it works

Clients connect over Streamable HTTP and discover the available MCP tools. The hosted service exposes 58 tools covering model discovery, sample generation, schema creation and editing, enrichment, job control, records, statistics, fusion, and database-related workflows.

A typical workflow starts by generating or reviewing sample JSON, creating a schema, and saving or publishing it. An agent can then enrich one entity synchronously or start asynchronous batch enrichment. Job status, persisted record IDs, costs, failures, and database outcomes can be inspected through the relevant tools. Existing records can be merged, and failed expertise areas can be retried without repeating successful work.

MCP resources provide an index and on-demand guides for schema modeling, documents, enrichment and fusion, batch jobs, benchmarks, database synchronization, and semantic identities. Reading these resources does not execute a model call.

Setup and configuration

The Entity Enricher MCP server needs no local installation. Configure the hosted endpoint in an MCP client using OAuth when the client supports the standard flow. The documented OAuth examples cover Claude Code, claude.ai, Cursor, and other OAuth-capable clients.

For clients using static JSON configuration, create an organization access key in the Entity Enricher web UI and send it as the X-API-Key header. Available roles include operator, editor, and owner; owner access is required for benchmarks. The key is shown only when it is created, so it must be stored by the operator.

Optional database delivery relies on ee-database running on the replica host. The database DSN remains on that host. Managed hosts may provision this arrangement automatically, while manual pairing is also supported.

Tools and capabilities

Notable tool groups include:

• Model discovery, including available model keys, languages, strategies, and organization limits.
• Schema generation from samples, direct schema saving, publishing, analysis, property edits, nesting, and soft deletion.
• Single-entity enrichment, asynchronous batch enrichment, record merging, fusion, and targeted retries.
• Job status checks, cancellation, and responses to questions from paused jobs.
• Record listing and retrieval, observed enum candidates, semantic identity work, and database schema or migration workflows.
• Benchmarks for enrichment, sample generation, and schema generation on the organization’s own tasks.

Limitations and notes

The Entity Enricher MCP server does not make schema validation or agreement between models proof that a result is factually correct or current. Review source material, failures, and partial outputs before relying on them.

A successful model generation, admission into the entity layer, and application to an external database replica are separate outcomes. Enrichment can be billed, and batch work is asynchronous. Tool descriptions expose preconditions, costs, and consequential effects, so clients should inspect those contracts before automating changes.

Client support depends on the authentication method. OAuth requires a client implementing the standard OAuth flow; static API-key configuration is intended for clients configured through JSON and for headless or CI use.