TeamCity MCP server

What the model can do

TeamCity CI/CD on your own server or TeamCity Cloud: builds by configuration, branch, state and status with who triggered them, one build with its test and problem counts, the build queue with wait reasons, build configurations and projects, failing tests and build problems, and build agents; when you ask, can run a build (optionally on a branch, with parameters), cancel a queued or running build, delete a build, comment on, tag, pin or unpin a build, assign an investigation, mute or unmute a test, enable or disable an agent, and pause or resume the whole build queue.

Reads

• teamcity_list_buildsBuilds in TeamCity, newest first: number, configuration, project, state (queued, running, finished), status (SUCCESS, FAILURE, ERROR), status text…
• teamcity_get_buildOne TeamCity build in detail: status and status text, branch, configuration, times, trigger, agent, tags, comment, whether it is pinned…
• teamcity_build_queueThe TeamCity build queue: queued builds with configuration, branch, position, why each is waiting, estimated start and who triggered it.
• teamcity_list_build_configurationsTeamCity build configurations (what can be run): id, name, project, whether paused, description and link.
• teamcity_list_projectsTeamCity projects: id, name, parent project, whether archived, description and link.
• teamcity_failed_testsFailing tests in TeamCity: for one build (its failed tests), or what is currently failing across a project.
• teamcity_build_problemsBuild problems in TeamCity (compilation errors, non-zero exit codes, failed steps): for one build, or what is currently failing in a project.
• teamcity_list_agentsTeamCity build agents: name, id, whether connected, enabled and authorized, up to date, version, pool, last activity and the build it is running.

Acts, when you ask

These change something in TeamCity. The model is told to use them only when you ask for exactly that, and each one is granted by name: per team on a shared connection, by the owner on a personal one.

• teamcity_run_buildQueues a new build of a TeamCity build configuration, on its default branch unless branch is given.
• teamcity_cancel_queued_buildCancels a TeamCity build that is still in the queue, so it never runs; it stays in history as canceled with your comment.
• teamcity_cancel_buildStops a running TeamCity build; it ends as canceled with your comment, and whatever it was doing (a deploy, a release) stops part-way.
• teamcity_delete_buildPermanently deletes a TeamCity build with its log, artifacts and test history; it cannot be recovered.
• teamcity_comment_buildSets the comment on a TeamCity build, replacing any comment it had; everyone who can see the build sees it.
• teamcity_tag_buildAdds tags to a TeamCity build (existing tags stay); tags are visible to everyone who can see the build.
• teamcity_pin_buildPins a TeamCity build (so clean-up never deletes it) or unpins it (pinned false), with an optional comment.
• teamcity_assign_investigationAssigns the investigation of a TeamCity build configuration's problems to a person (by TeamCity username), who is notified.
• teamcity_mute_testMutes a test in a TeamCity project, so its failures no longer fail builds there, until it is fixed (default) or unmuted by hand.
• teamcity_unmute_testRemoves a TeamCity mute by its id, so the muted test's failures fail builds again.
• teamcity_set_agent_enabledEnables or disables a TeamCity build agent.
• teamcity_pause_queuePauses (paused true) or resumes the whole TeamCity build queue: while paused no queued build starts on any agent, for every project.

Connect it

First connect TeamCity on Dashboard → Connections, then add Memtro to your client once. The TeamCity tools appear alongside the tools of everything else you have connected; nothing is configured per connector.

claude plugin marketplace add memtro-org/claude-memtro
claude plugin install memtro@memtro --config api_key=sk-memtro-...

{
  "mcpServers": {
    "memtro": {
      "url": "https://platform.memtro.com/api/mcp",
      "headers": {
        "Authorization": "Bearer sk-memtro-..."
      }
    }
  }
}

{
  "mcp": {
    "memtro": {
      "type": "remote",
      "url": "https://platform.memtro.com/api/mcp",
      "headers": {
        "Authorization": "Bearer sk-memtro-..."
      }
    }
  }
}

Who may use it

The key identifies the person, and the server offers that person the tools they may use everywhere else in Memtro. A shared TeamCity connection is used by the teams an administrator granted it to, tool by tool, on the Permissions page; a personal connection is used by its owner alone, and its tools that act stay off until the owner allows them. See Teams and permissions and Connectors.

Ask it like this

• › Why did the last main build of the web project fail?
• › What is waiting in the build queue, and why?
• › Which tests are currently failing in the API project?
• › Which build agents are disconnected?
• › Run the nightly build of the web project on the release branch.

Questions

Is this an official TeamCity MCP server?

No. It is Memtro's MCP server with Memtro's TeamCity connector behind it: the connector reads TeamCity through its API with the credentials you connect on Dashboard → Connections, and TeamCity publishes nothing here. The same endpoint carries every other system you connect and Memtro's memory tools.

What does it cost?

Nothing on top of Memtro. One price per organisation by its number of users, from $50 a month for up to 10 users after a 7-day free trial, with every connector, the MCP server, jobs and memory included. You bring your own model keys.

Can it change anything in TeamCity?

Only when you ask, and only what it has been granted: run a build of a configuration, optionally on a branch or with parameters; cancel a queued build; cancel a running build; delete a finished build with its artifacts and history (permanent); comment on a build; tag a build; pin or unpin a build; assign an investigation of a configuration's problems to a person; mute a test in a project; unmute a test; enable or disable a build agent; pause or resume the whole build queue. Every other tool reads, searches or lists. An administrator grants each acting tool by name per team, and a personal connection acts only after its owner allows it.

Which clients work?

Any MCP client that speaks Streamable HTTP: Claude Code, Cursor, Claude Desktop, OpenCode, Codex, Zed, Windsurf, Cline and the Anthropic and OpenAI SDKs. There is no local process to run; the server is remote and authenticated with your key.

Other MCP servers · Engineering and operations

• GitHub MCP server
• AWS CloudWatch MCP server
• Sentry MCP server
• ServiceNow MCP server
• Okta MCP server
• Twilio MCP server