mcp · teamwork — Claude Code MCP server: what it does, how to install
Teamwork.com's official MCP server hands your agent projects, tasks and tickets, and lets it change them on request. Useful when the work lives outside your editor.
claude mcp add --transport http teamwork-mcp https://mcp.ai.teamwork.com/sseWhat mcp does
The server spells out Teamwork.com actions in a shape an agent can pick up and call on its own: projects, tasks and timers, Desk tickets, Spaces pages, Chat conversations. Point a client at the hosted endpoint https://mcp.ai.teamwork.com/sse, or run the container docker.io/teamwork/mcp:v1.51.3 over stdio when you want it on your own machine.
Under the hood there is a Go project with three entry points: an HTTP server for several clients at once, a stdio server for desktop tools, and a CLI that prints the tool list for debugging. Access comes from a Teamwork.com bearer token or OAuth2. A read-only switch is there too, and it is the setting to pick while you are still watching what the agent does with write calls.
Building from source takes Go 1.27 or newer. The stdio container reads TW_MCP_BEARER_TOKEN from the environment; leave it empty and the API refuses you. Debugging OAuth2 against Let's Encrypt also needs NODE_EXTRA_CA_CERTS, otherwise the certificate check stops MCP Inspector cold.
What it helps with
- list a project's tasks with tags and running timers
- clear a backlog of unanswered Desk tickets
- locate a Spaces page by its title
- summarise a Chat thread and pull out the decisions
- file a task straight from a conversation
Example prompts for the agent
- Show the Teamwork tasks assigned to me this week
- Create a task in the Redesign project due Friday
- Which Desk tickets still have no reply?
How to install mcp
1Platform CLI: remote server
claude mcp add --transport http teamwork-mcp https://mcp.ai.teamwork.com/sse2Platform CLI
claude mcp add --env TW_MCP_BEARER_TOKEN=<TW_MCP_BEARER_TOKEN> --env TW_MCP_LOG_FORMAT=<TW_MCP_LOG_FORMAT> --env TW_MCP_LOG_LEVEL=<TW_MCP_LOG_LEVEL> --transport stdio teamwork-mcp -- docker run -i --rm -e TW_MCP_BEARER_TOKEN -e TW_MCP_LOG_FORMAT -e TW_MCP_LOG_LEVEL docker.io/teamwork/mcp:v1.51.3Replace <…> with your own values
1Platform CLI: remote server
codex mcp add teamwork-mcp --url https://mcp.ai.teamwork.com/sse2Platform CLI
codex mcp add teamwork-mcp --env TW_MCP_BEARER_TOKEN=<TW_MCP_BEARER_TOKEN> --env TW_MCP_LOG_FORMAT=<TW_MCP_LOG_FORMAT> --env TW_MCP_LOG_LEVEL=<TW_MCP_LOG_LEVEL> -- docker run -i --rm -e TW_MCP_BEARER_TOKEN -e TW_MCP_LOG_FORMAT -e TW_MCP_LOG_LEVEL docker.io/teamwork/mcp:v1.51.3Replace <…> with your own values
Config file
{
"mcpServers": {
"teamwork-mcp": {
"url": "https://mcp.ai.teamwork.com/sse"
}
}
}{
"mcpServers": {
"teamwork-mcp": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"-e",
"TW_MCP_BEARER_TOKEN",
"-e",
"TW_MCP_LOG_FORMAT",
"-e",
"TW_MCP_LOG_LEVEL",
"docker.io/teamwork/mcp:v1.51.3"
],
"env": {
"TW_MCP_BEARER_TOKEN": "<TW_MCP_BEARER_TOKEN>",
"TW_MCP_LOG_FORMAT": "<TW_MCP_LOG_FORMAT>",
"TW_MCP_LOG_LEVEL": "<TW_MCP_LOG_LEVEL>"
}
}
}
}Replace <…> with your own values
Config file
{
"servers": {
"teamwork-mcp": {
"type": "http",
"url": "https://mcp.ai.teamwork.com/sse"
}
}
}{
"servers": {
"teamwork-mcp": {
"type": "stdio",
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"-e",
"TW_MCP_BEARER_TOKEN",
"-e",
"TW_MCP_LOG_FORMAT",
"-e",
"TW_MCP_LOG_LEVEL",
"docker.io/teamwork/mcp:v1.51.3"
],
"env": {
"TW_MCP_BEARER_TOKEN": "<TW_MCP_BEARER_TOKEN>",
"TW_MCP_LOG_FORMAT": "<TW_MCP_LOG_FORMAT>",
"TW_MCP_LOG_LEVEL": "<TW_MCP_LOG_LEVEL>"
}
}
}
}Replace <…> with your own values
1Platform CLI: remote server
gemini mcp add -t http teamwork-mcp https://mcp.ai.teamwork.com/sse2Platform CLI
gemini mcp add -e TW_MCP_BEARER_TOKEN=<TW_MCP_BEARER_TOKEN> -e TW_MCP_LOG_FORMAT=<TW_MCP_LOG_FORMAT> -e TW_MCP_LOG_LEVEL=<TW_MCP_LOG_LEVEL> teamwork-mcp docker run -i --rm -e TW_MCP_BEARER_TOKEN -e TW_MCP_LOG_FORMAT -e TW_MCP_LOG_LEVEL docker.io/teamwork/mcp:v1.51.3Replace <…> with your own values
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about mcp
How do I install mcp in Claude Code?
Register the remote endpoint with one command: claude mcp add --transport http teamwork-mcp https://mcp.ai.teamwork.com/sse. The client then lists the server as teamwork-mcp.
What does mcp do?
It exposes Teamwork.com tools to the agent: projects, tasks, Desk tickets, Spaces pages and Chat messages. The agent calls them when you ask it to find, create or update something.
Does mcp need a token to work?
Yes, a Teamwork.com bearer token or a working OAuth2 setup. The stdio container picks the token up from TW_MCP_BEARER_TOKEN.
Similar MCP servers
All MCP servers for Agent setup: memory, context, orchestration →mcp by other authors comes first.
