Provides the capability to use the CLAIRE Orchestrator MCP server. When you add this MCP server, IDMC orchestrates the full data management flow, including discovery, data management, data quality, and other data management skills.
The CLAIRE Orchestrator MCP server is the primary entry point into Informatica CLAIRE. When you give it a request in a natural language, it plans the end-to-end data management workflow and routes work to specialized downstream skills, including metadata discovery, data exploration, data quality analysis, rule recommendation and generation, cleansing, profiling, product enrichment, data integration, and MDM.
Key capabilities
This MCP server offers the following capabilities:
•Understanding and planning. Translates natural-language requests into multi-step plans across IDMC data management.
•Conversation state management. Maintains multi-turn conversation context and resolves references and asset IDs across turns.
•Skill routing. Routes requests to the appropriate downstream CLAIRE skill.
Use cases
Use this MCP server to address the following use cases:
•Unified platform access. Enable AI agents to use the entire IDMC platform through a single MCP server without needing separate connections for each service.
•Workflow delegation. Allow AI agents to delegate long-running data management workflows to the orchestrator without hard-coding IDMC skill boundaries.
•Natural-language task sequencing. Allow users to request an outcome, such as "Clean up my customer data," and have the orchestrator automatically sequence the workflow, including discovery, profiling, and cleansing.
Tools
The following table describes the tools available on this MCP server:
Tools
Description
Get_skill_execution_updates
STREAMING RENDER LOOP. This tool returns one part of a CLAIRE response. Everything in 'items[]' is finished, user-facing text that CLAIRE has already produced. Render it to the user BEFORE calling this tool again. DO NOT accumulate items across polls and write a single summary at the end or the tool will fail. The user is waiting on each slice; a correct final answer delivered all at once is still wrong. Loop Call this after invoking a skill with 'deferred: true' that returned 'status: "IN_PROGRESS"'.
1Render every item in 'items[]' per the contract below.
3Call again with 'pagination.cursor' passed back verbatim; keep 'message_id and conversation_id' unchanged.
4Stop when 'status' is terminal (SUCCESS | PARTIAL_SUCCESS | FAILED) AND 'pagination.has_more' is false.
On the first call, 'message_id and conversation_id' comes from 'context' in the initial response, and 'cursor' is "0". An empty 'items[]' is normal. Poll again. 'Render_note' is only present when there is output to emit. Do not call other tools between polls unless the user asks something new.
Rendering contract (mandatory). Render in arrival order. Never summarize, paraphrase, translate, reorder, omit, restructure, rename, or truncate a frame.
- TOKEN Buffer 'payload.content' between the enclosing STREAM_STATUS STARTED and COMPLETED; emit the concatenated text as one block on COMPLETED. Never emit individual fragments. A run may span polls — keep the buffer across calls.
- STREAM_STATUS Not rendered; drives the TOKEN flush.
- HUMAN_PROBE Verbatim. Stop polling and wait for the user's reply.
- PLAN Plan text verbatim, then list 'suggestedPrompts'.
- ARTIFACT Render each entry of 'payload.data[]' according to 'viewType':
- GridView Markdown table with EVERY column in 'schema_definitions' and EVERY row in 'items'. Use each column's 'label' as the header, unchanged. Do not drop columns that look internal, do not shorten cell values, do not split one grid into multiple tables, and do not add columns.
- BlockText Text content only. Strip HTML tags and custom elements such as <infa-grid>; render any text they wrap. If a frame is only markup with no text, render nothing for it.
- PropertiesView Key/value list.
- TreeView Indented tree.
- GraphView Source -> target list.
- DataExploreResult SQL in a code block, then rows.
- AnalyticsView Metric list.
Orchestrate
Interprets user intent and routes queries across IDMC capabilities, including metadata discovery, data exploration, data quality analysis, rule recommendation and generation, cleansing, profiling, product enrichment, data integration, MDM, and product documentation. Resolves references from conversation history and dispatches sub-queries to the correct downstream skill. Use this action for general CLAIRE conversations where the appropriate capability must be selected based on the user's request.
This action is an agent-backed MCP server endpoint. For correct routing, pass the user prompt verbatim in the prompt argument. Do not summarize the prompt.
Orchestrate_master_data_analysis
Queries about mastered entities, business entities, or MDM relationship and hierarchy questions. Coordinates downstream MDM skills for resolved entities, golden records, hierarchies, match/merge contexts, and cross-entity relationships. Use this action to handle requests from the MDM context and route queries across MDM capabilities, including grounding and ambiguity resolution, MDM relationship exploration, and metadata discovery and retrieval.
This action is an agent-backed MCP server endpoint. For correct routing, pass the user prompt verbatim in the prompt argument. Do not summarize the prompt.
Request_upload_url
Returns a single-use PUT URL that's valid for 5 minutes. Call this tool first, then upload the file using the returned curl command. If conversation_id is omitted, a new conversation is created. If the PUT request returns HTTP 401, call this tool again to create a new URL and try again.
Note:
In AI Agent Engineering, the MCP server tools are called "actions."