Fireflies.ai's official remote MCP server — 19 tools for searching meeting transcripts and summaries, pulling action items and analytics, and cutting soundbites from recorded calls.
Teamwork.com MCP
Summary
Teamwork.com's official MCP server for projects, tasks, time, Desk tickets, Spaces and Chat — with profile endpoints that trim the tool list to one role and an optional read-only mode.
Features
- Projects, tasks, time tracking, Desk tickets, Spaces and Chat in one server
- Five profile endpoints scope the tool list to a role instead of loading everything
- Optional read-only mode for agents that should look but not touch
- HTTP and stdio transports with bearer-token or OAuth2 authentication
- Hosted endpoint at mcp.ai.teamwork.com — no self-hosting required
- Auto-generated, searchable tool reference published by the maintainers
Installation
Set up this MCP server in your favorite AI agent — copy a ready-made configuration below.
Description
Teamwork.com's official MCP server exposes its project-management, help-desk and knowledge surfaces to an agent: Projects, Desk, Spaces and Chat, across create, read and update operations. It is written in Go and open source under the MIT licence.
Profile endpoints
The design choice worth noting is that the server does not force one enormous tool list on every client. The main hosted endpoint at https://mcp.ai.teamwork.com loads everything, but five profile endpoints scope it to a role:
/project-manager— projects, tasks, people and content/support(aliased as/desk) — help-desk tickets and customers/analyst— projects, tasks, people, time tracking, content, tickets, customers and desk admin/knowledge-manager— spaces, pages and content/ops— everything
A smaller tool list is a real quality lever for agents: fewer near-duplicate tools means fewer wrong picks. All profiles accept the same authentication.
Deployment and safety
Both HTTP and stdio transports are supported, with bearer-token or OAuth2 authentication against Teamwork.com. A read-only mode restricts the server to reads, which is the setting you want when an agent is exploring rather than acting. The project ships production logging and observability, and an auto-generated tool reference is browsable at teamwork.github.io/mcp.
Getting started
An administrator enables MCP for the site under Settings → AI, and you need permission to create an API key. Client guides cover the Teamwork CLI, ChatGPT, Claude Desktop, Claude Code, VS Code with Copilot Chat, Gemini CLI, n8n and Appmixer. Building from source needs Go 1.27 or later.
Related MCP Servers
ClickUp's official remote MCP server — around 40 tools across tasks, search, Docs, Chat and time tracking, over OAuth 2.1 with PKCE.
Metabase's built-in MCP server — search, query and visualise your BI data through the semantic layer, scoped to the connecting person's existing Metabase permissions.
Zoom's official hosted MCP server — semantic meeting search, AI Companion summaries, transcripts, recordings, and cross-Zoom search over Team Chat, Canvas and My Notes.