Coveo Launches Hosted MCP Server for Enterprise AI Integration

Coveo Solutions Inc. has launched a hosted Model Context Protocol (MCP) Server to expand interoperability between its AI-Relevance Platform and leading large language models, announced in a press release. The new server enables secure and direct access for AI agents to enterprise content indexed by Coveo without requiring custom integrations or additional infrastructure.

Built on the MCP standard, the hosted server connects Coveo’s unified content index with models such as ChatGPT Enterprise and Claude by Anthropic. This allows organizations to link multiple data sources to large language models while maintaining enterprise security and governance.

Coveo’s app is also now available in ChatGPT Enterprise’s official Apps & Connectors directory, providing a secure bridge between ChatGPT and Coveo-indexed content. The integration supports grounded retrieval, reuse of Coveo’s relevance logic, and deployment across various AI-driven use cases.

According to the company, ten customers are already using the hosted MCP Server to enhance models from Anthropic and OpenAI. The MCP Server is generally available as part of the Coveo AI-Relevance Platform under existing consumption-based licensing.