Ask Senzing

Ask anything about Senzing — products, the SDK, entity-resolution concepts, configuration, and error codes — and it politely declines everything else.

Click the floating ✦ Ask Senzing button in the bottom-right corner to start a chat. Ask a real Senzing question and watch it search the docs live.

Things You Can Ask

"How do I load JSON data with the Python SDK?"
"What does error SENZ0005 mean?"

Bring Your Own LLM

Use the Senzing MCP with your own LLM

The same Senzing knowledge that powers this assistant is available as a public MCP server. Connect it to your own LLM or agent for accurate, up-to-date Senzing answers grounded in the same source of truth.

MCP Endpoint

https://mcp.senzing.com/mcp

To get started, ask your AI assistant:
"Add the Senzing MCP server at https://mcp.senzing.com/mcp"

Streamable HTTP transport. It works via the Anthropic Messages API mcp_servers connector and any MCP-aware client (Claude Desktop and other MCP clients). No authentication required.

What you get

get_capabilitiesDiscover the full toolset and workflows first

search_docsPrimary docs / pricing / architecture / deployment retrieval

explain_error_codeLook up Senzing error codes and meanings

find_examplesWorking code from indexed GitHub repos

get_sdk_referenceV4 method signatures, flags, V3→V4 migration

sdk_guideInstall / configure / load / export across platforms

generate_scaffoldGenerate correct SDK initialization & scaffolding

mapping_workflowMap source JSON to the Senzing entity model

reporting_guideER reporting & visualization patterns

Add it as an MCP server

{
  "mcpServers": {
    "senzing": { "type": "url", "url": "https://mcp.senzing.com/mcp" }
  }
}

Tool-first rule. Tell your LLM to call get_capabilities first, and to treat the Senzing MCP as the exclusive source of truth for anything Senzing.

Training data is often wrong or outdated for Senzing — wrong attribute names, wrong method signatures, wrong base images — so the model should use the MCP's tools and never supplement or "improve" their results with prior knowledge.

How it works

Every answer is grounded in the Senzing MCP as the single source of truth. There is no cached or derived index — the assistant produces each answer from live MCP calls made at the moment you ask.

The model drives the MCP's tools agentically: it chooses which tools to call (search, error-code lookup, code examples, SDK reference), refines and chains those calls, then synthesizes a grounded answer that cites the docs.

Scope is structural, not just prompt-based. The assistant can reach nothing but the Senzing MCP — no web search, no other data source. An off-topic question surfaces nothing relevant, so the assistant declines and points you back to what it can help with.

Notice

Provided AS-IS, with no warranty of any kind — no warranty of merchantability, fitness for a particular purpose, accuracy, completeness, or reliability. Use entirely at your own risk, and verify any output independently before relying on it for production decisions.

Do not send PII, credentials, or sensitive data. Questions and parameters may be logged. This assistant answers documentation questions; it does not process your data and cannot perform actual entity resolution.