Skip to main content
Technical reference for the Credible Data MCP server, available at https://<org>.mcp.credibledata.com/mcp.

​
Core Tools

​
suggestAnalysis

Translates natural language requests into executable Malloy queries based on your semantic data models. Parameters:
  • naturalLanguageQuery (required): The user’s question in natural language (e.g., “What were our top-selling products last year?”)
  • projectName (optional): Project name to search within. Only use if known from context or explicitly mentioned by user.
  • packageName (optional): Package name to narrow search scope. Requires projectName.
  • modelPath (optional): Path to specific .malloy model file (e.g., models/sales_model.malloy). Requires projectName and packageName.
  • sourceName (optional): Specific source/table within a model. Requires projectName, packageName, and modelPath.
Parameter Dependencies: projectName → packageName → modelPath → sourceName Scope Strategy: Start broad when uncertain, narrow as you discover structure. If results are insufficient, widen scope by removing parameters from right to left. Response:
  • matchingResults: Array of query suggestions with executable Malloy query code and primaryTransactionTable
  • entities: Available dimensions, measures, views, and columns you can use
  • docs: Malloy syntax documentation and common errors to avoid
  • nextSteps: Actionable guidance for refining analysis
Example Request: 
curl -X POST "https://your-org.mcp.credibledata.com/mcp" \
  -H "Content-Type: application/json" \
  -H "Authorization: ApiKey YOUR_API_KEY" \
  -d '{
    "jsonrpc": "2.0",
    "id": 1,
    "method": "tools/call",
    "params": {
      "name": "suggestAnalysis",
      "arguments": {
        "naturalLanguageQuery": "What are the top 10 products by sales?",
        "projectName": "your-project"
      }
    }
  }'

​
executeQuery

Executes Malloy queries against data models and returns JSON results. Parameters:
  • projectName (required): Project containing the model
  • packageName (required): Package containing the model
  • modelPath (required): Path to the .malloy model file
  • query (optional)*: Custom Malloy query code. Do NOT provide sourceName when using this.
  • queryName (optional)*: Name of predefined query/view to execute
  • sourceName (optional)*: Source name. Required when using queryName, omit when using custom query.
*Execution Patterns: Use exactly ONE of:
  1. Custom query: Provide query parameter only
  2. Predefined query: Provide both queryName and sourceName
Response: Returns query results as JSON with data, totalRows, executionTime, and metadata

​
Discovery Tools

Deprecated. These tools are only available via the legacy https://<org>.data.credibledata.com/mcp endpoint and will be removed in a future release. Use suggestAnalysis instead.
Navigate the organizational hierarchy to discover available data before analysis. Tool Hierarchy: listProjects → listPackages → listModels → getModel

​
listProjects

Lists all projects in your organization. Parameters: None (organization derived from auth context) Returns: Array of projects with name and description

​
getProject

Gets details of a specific project. Parameters: projectName (required)

​
listPackages

Lists all packages within a project. Parameters: projectName (required) Returns: Array of packages with name, description, and metadata

​
getPackage

Gets detailed package information including contents. Parameters: projectName (required), packageName (required)

​
listModels

Lists all .malloy model files and .malloynb notebooks in a package. Parameters: projectName (required), packageName (required) Returns: Array of models with path, type, and metadata

​
getModel

Gets compiled model content including sources, views, queries, measures, and dimensions. Parameters: projectName (required), packageName (required), path (required)

​
Authentication

The MCP server supports two authentication methods:
  • OAuth 2.0: For connecting to AI assistants like Claude Desktop. See AI Assistants with MCP.
  • API Key: For building your own custom agents with server-to-server communication. Pass as Authorization: ApiKey YOUR_API_KEY header. See Custom AI Agents.
Configure your MCP client with your organization’s endpoint: https://<org>.mcp.credibledata.com/mcp

​
Error Handling

The server returns standard MCP error responses for invalid requests, authentication failures, and query execution errors. Refer to the MCP specification for error code details.