Connect AI Agents to trusted enterprise context

MCP is an open standard that gives AI agents and large language models a consistent way to connect to external systems. It lets an AI assistant or agent securely reach an application, such as a data catalog, its data sources, and its tools, so the AI’s answers and actions stay grounded in your organization’s real data, not just its training data.

The Actian MCP Server is a secure bridge between your AI tools, such as Claude or Microsoft Copilot, and the metadata inside the Actian Data Intelligence Platform. Human users and AI agents can both query the catalog to find data, understand business terms, and map data relationships, all without leaving their workflow.

Both. Agents can also set or restore a description, regenerate a summary, or set a custom property. Every write requires explicit confirmation before it happens, so agents propose changes and a person still approves them.

Every write carries an application identifier and is logged to the same audit trail as changes made by people. You can see who, or which agent, changed what, and when, at any time. Nothing an agent writes is quietly invisible to your governance team.

The Actian MCP Server supports OAuth, so agents authenticate through your existing identity provider. You control which agents can connect and what they can do, using the access controls you already have.

• Knowledge Graph foundation. Actian Data Intelligence Platform is built on a native knowledge graph, so the MCP Server delivers a more accurate understanding of complex data relationships and semantic meaning than platforms that are not knowledge graph native.
• Open standards, not lock-in. Building on the open MCP standard means the server is designed for interoperability and evolves with the AI ecosystem, instead of tying you to one vendor’s assistant.

Confirmed compatible tools include Claude, Cursor, Windsurf, Microsoft Copilot Studio, and ChatGPT. Because the server is built on the open MCP standard, we expect this list to grow as the ecosystem does.

• Look Up Official Definitions: Use the find_glossary_definition tool to retrieve the exact, approved definition for any business term stored in your glossary.
• Find Data Behind a Concept: Use the get_glossary_implementations tool to locate all the physical data assets (tables, files, etc.) that are linked to a specific business concept.
• Map Data Relationships: Use the get_dataset_data_model tool to instantly understand how a dataset is connected to others, including its specific fields and relational dependencies.
• Discover Related Ideas: Use the semantic_search_glossary tool to perform an AI-powered search that finds conceptually similar business terms, even if they don’t share the same keywords.
• Documentation Backfill: The agent writes descriptions for undocumented items with update_item_description, or puts an item back to the description from the source system.
• Summary Upkeep: update_item_summary sets a summary or regenerates it from the description, and explains why if it cannot.
• Property Curation: update_item_property sets or clears properties such as domain, criticality, tags, URLs, or dates, finding the property by ID, code, or name and matching the value to its type.
• Impact Analysis: fetch_lineage returns the lineage from an item, so the agent can say what a change would affect upstream and downstream.
• Audit Questions: get_item_audit_events shows what changed on an item, who or what changed it, and when.