It is a thin bridge rather than a reimplementation: results come from your organization's index and the answers come from Glean's own chatbot, so permissions and coverage stay whatever Glean already gives that key. Setup is a docker build and two environment variables.
An MCP server over the Glean API, exposing Glean's two main surfaces: search across your indexed workplace content, and Q&A with Glean's chatbot.
- Search — a list of results for a query, from your organization's Glean index
- Chat — question and answer with the Glean chatbot
- Ships as a Dockerfile, built from src/glean/Dockerfile
GLEAN_API_KEY and GLEAN_DOMAIN, passed into the container by your client's config. Build the image yourself with docker build -t glean-server:latest -f src/glean/Dockerfile . — there is no published package.
One command plus a key — docker build -t glean-server:latest -f src/glean/Dockerfile ., then supply credentials
