Glean MCP Server
# Glean Local MCP Server Monorepo
> [!IMPORTANT]
> This repository is archived / no longer maintained.
>
> Do not use this repository for new MCP setups. It predates Glean’s managed remote MCP server and is no longer the supported path for connecting MCP-compatible hosts to Glean.
## What to use instead
Use the managed Glean MCP server built into your Glean instance:
- [Using the Glean MCP Server](https://docs.glean.com/user-guide/mcp/usage) — end-user setup for supported MCP hosts.
- [Set up Glean MCP server](https://docs.glean.com/administration/platform/mcp/enable-mcp-servers) — admin setup and enablement.
- [About Glean MCP server](https://docs.glean.com/administration/platform/mcp/about) — overview of Glean’s managed MCP server.
For local command-line workflows, use [Glean CLI](https://github.com/gleanwork/glean-cli) instead of this local MCP server.
## Repository status
This repository is retained for historical reference only. Issues and pull requests have been closed as part of archival cleanup, and no new feature work or bug fixes are planned here.
The local packages in this repository should not be used for new installations. Published user-facing local MCP packages are being retired in favor of the managed Glean MCP server.
## What this repository used to contain
This monorepo previously contained packages for running and supporting a local stdio-based MCP server for Glean:
- `@gleanwork/local-mcp-server` — a local MCP server that exposed Glean Search, Chat, People Search, and document-reading tools.
- `@gleanwork/mcp-server-utils` — shared utilities used by the local MCP server packages.
These packages predate the managed remote MCP server and are no longer the recommended or supported MCP integration path.
## Historical reference
Existing install, Docker, and client-configuration instructions have intentionally been removed from the main documentation so new users are not directed toward an unsupported setup. If you need to understand the old implementation, inspect the repository history.
## License
MIT License — see the [LICENSE](LICENSE) file for details.
TDQS
Scored across 3 tools
Each tool has a clearly distinct purpose: 'chat' is for conversational interaction with an assistant, 'company_search' is for finding documents and data, and 'people_profile_search' is for locating employee profiles. There is no overlap in functionality, making it easy for an agent to select the correct tool based on the task.
The tool names follow a consistent snake_case pattern and are descriptive, but there is a minor deviation: 'chat' uses a simple verb, while the other two tools use a noun_noun structure (e.g., 'company_search'). This slight inconsistency does not hinder readability or predictability significantly.
With only 3 tools, the set feels thin for a server named 'Glean MCP Server', which implies broader capabilities in information retrieval and assistance. While the tools cover chat, document search, and people search, the scope might benefit from additional tools for more granular operations or updates, making the count borderline appropriate.
The tools provide basic search and chat functionalities, but there are notable gaps. For example, there is no tool for updating or managing data (e.g., creating documents or modifying profiles), and the chat tool lacks explicit support for follow-up actions or context persistence. This limits the server's ability to handle full lifecycle operations in its domain.