This section provides details on the fields required to configure an AI Catalog datasource in Data Integrity Suite.
Datasource configuration
When you select the AI Catalog tile, the following fields appear before proceeding to the connection wizard.
| Field | Description |
|---|---|
| Datasource name | Specifies the display name for the datasource. This is a mandatory field. Example: My AI Catalog |
| Description | Allows you to describe the purpose of this datasource. Example: Catalog of internal AI agents and MCP servers. |
Connection configuration
The following fields are available in the connection wizard. Fields marked as MCP / AI Agent mode only appear when MCP / AI Agent is selected as the Discovery Source.
| Field | Description |
|---|---|
| Connection name | Specifies the name of the data connection. This is a mandatory field. Example: Default |
| Description | Allows you to describe the purpose of this connection. Example: Connection to internal MCP servers and AI agents. |
| Agent to access data | Select the Precisely agent from the dropdown that connects to your network. For cloud AI providers where data is accessed directly over the internet, select None (data is in the cloud). |
| Discovery Source | Specifies the type of AI assets to discover. Select one of:
|
| Organisation Namespace (MCP / AI Agent mode only) | Specifies the namespace used to identify your organization's AI platform in the catalog. This value groups all discovered assets under a single AI Platform Registry. Example: io.github.myorg |
| Platform Name (MCP / AI Agent mode only) | Specifies the display name for the AI platform registry created in the catalog. Example: My Org AI Platform |
| MCP Server Endpoints (MCP / AI Agent mode only) | Comma-separated list of MCP server endpoint URLs to introspect. The connector calls each endpoint using JSON-RPC to discover tools, resources, and prompts. Example: http://mcp.myorg.com/server, http://mcp.myorg.com/tools |
| AI Agent Endpoints (MCP / AI Agent mode only) | Comma-separated list of AI agent base URLs. The connector fetches the AgentCard from each URL's well-known path to discover agents and their skills. Example: http://agent.myorg.com:9090, http://agents.myorg.com/.well-known/agents |
Cloud AI Provider configuration
When Cloud AI Providers is selected as the Discovery Source, the following provider-specific fields are available depending on the provider chosen.
| Provider | Field | Description |
|---|---|---|
| Anthropic Claude | Anthropic API Key | API key for your Anthropic account. Must start with
sk-ant-. Generate at
console.anthropic.com. |
| OpenAI | OpenAI API Key | API key for your OpenAI account. Must start with
sk-. Generate at
platform.openai.com. |
| GitHub Copilot | GitHub Organisation | The GitHub organization name to discover Copilot assets for. Example: myorg |
| GitHub Token | A classic personal access token (PAT) with repo
and read:org scopes. Fine-grained PATs are not
supported for organization-level Copilot APIs. |
|
| Microsoft Copilot Studio | Azure Tenant ID | The Directory (tenant) ID from Microsoft Entra ID Overview. Example: xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx |
| Azure Client ID | The Application (client) ID from the Azure App Registration. Example: xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx | |
| Azure Client Secret | The secret value generated under Certificates & secrets in the Azure App Registration. |
Options
| Field | Description |
|---|---|
| Request Timeout (seconds) | Maximum time in seconds to wait for a response from MCP server endpoints or agent card URLs. Default is 30. Increase this value for slow or remote servers. |
| Max Servers Per Harvest Run | Maximum number of MCP servers to introspect in a single harvest run. Default is 100. Set to 0 for unlimited. |
| Log Level | Sets the verbosity of connector logs. Options are Error, Warning, Info, and Debug. Default is Error. |