)
Connect Teamwork.com MCP to your AI agents, securely.
One secured endpoint for every Teamwork.com call your AI agents make, scoped per user, inspected, and logged.Try Teamwork.com MCP now
What your agents can ask once Teamwork.com MCP is connected.
Select the tools your AI agents can call
225 tools from the Teamwork.com MCP schema.
Showing 225 tools
What Teamwork.com MCP exposes
Teamwork.com MCP groups its 225 tools into 14 categories, each with a suggested starting policy.
All 225 tools as JSON, ready to paste.
| Category | Covers | Access | Suggested policy |
|---|---|---|---|
| Projects — Projects | Projects, categories, templates and members, custom fields and their values singly or in bulk, and custom item types such as Contracts, Leads or Deals with their fields and records, plus file upload into a project's files area and reserved upload URLs | Read / Write | add_project_member grants a person access to everything in a project, so it is an entitlement change wearing a membership name. create_upload_url reserves a short-lived URL that bytes are sent to directly, so the transfer itself never passes through the server and is invisible to call-level logging |
| Projects — Content | Comments, messages and replies, milestones, notebooks, tags and links, the project files area with file details and content download, the recent activity feed, and a cross-entity keyword search spanning projects, tasks, messages, comments and users | Read / Write | download_file returns the stored content itself, with text as text, images as images and anything else base64, so it is a content ingress path rather than a metadata read. The cross-entity search reaches everything the authorising account can see in one call |
| Projects — Tasks | Tasks, tasklists, workflows and workflow stages, including creating and updating many tasks in one call, completing and reopening tasks, moving tasks with their subtasks to another tasklist, and moving tasks through workflow stages | Read / Write | create_tasks and update_tasks act on many records per call, so one approval covers an unbounded set. move_tasks carries subtasks with it and preserves the parent and child structure, which makes it broader than it reads |
| Projects — People | Users, companies or clients, teams, skills and job roles, including creating a user, setting or clearing the job role of one or more users at a time, and reading the currently authenticated user | Read / Write | create_user adds a person to the installation, which is an identity the rest of the product will trust. set_user_jobrole and clear_user_jobrole both accept several users at once, so the scope of a single call is worth logging |
| Projects — Time | Timelogs and timers including creating, pausing, resuming and converting a timer to a timelog, calendars and time-blocking events, project and tasklist budgets, and complete time totals grouped by user, project, task or period | Read / Write | Timelogs are what clients are billed from, so creating or updating one changes a financial record rather than a note. The budget reads and the summarize tool together give a clear picture of commercial exposure per project |
| Projects — Planning | Scheduler allocations that commit a person's time to a project over a date range, links between tasks and allocations, restoring a deleted allocation, custom per-day capacity splits for a task assignee, and the workload view of planned task time | Read / Write | Allocations and workload are separate planes of planning and are not summed, which is easy to get wrong when reporting. restore_allocation is the only tool on the server that brings back a deleted record, since no delete tool is published |
| Desk — Admin | Inbox configuration in Teamwork Desk: ticket priorities, statuses, types and tags with full create, read and update, plus read-only listings of custom fields with their kind, ticket sources and the happiness ratings a customer can give | Read / Write | These define the vocabulary every ticket is classified against, so renaming a status or priority changes what every report and filter means without touching a ticket. Nothing here reaches a customer directly |
| Desk — Tickets | Creating, reading, updating and searching tickets, listing and reading inboxes, uploading files, and linking or unlinking a Teamwork Projects task to a ticket so the ticket shows the work behind it | Read / Write | reply_ticket writes to the customer on the other end of the ticket, which makes it the outbound act of this toolset and the one to gate hardest. Ticket reads carry whatever customers have written, so treat the content as untrusted text |
| Desk — Customers | Customer organizations, individual customers and support agents, with create and update on companies and customers and read-only access to agents | Read / Write | These records are personal data about people outside the organization. Agents are read-only here, which keeps the agent roster out of reach of an automated change |
| Desk — Helpdocs | Help doc articles with an HTML body, including creating and updating one, searching articles, listing categories, and reading the knowledge base sites with their subdomain, branding and article counts | Read / Write | A help doc site is customer-facing and has its own subdomain, so creating or updating an article publishes agent-written HTML to the public. Treat both writes as publication rather than as content editing |
| Spaces — Content | Comments on pages, Spaces tags created singly or several at a time, space categories, and full-text search across pages | Read / Write | Comments are visible to everyone with access to the page, so they are communication rather than a record change. The full-text search spans page contents the authorising account can reach |
| Chat | Reading conversations the current user belongs to, their messages and the people in the installation, reading the authenticated user with unread and mention counts, resolving or creating a one-to-one conversation, and sending a message to a conversation or a direct message to a person | Read / Write | send_message and send_dm put words in the user's name in front of colleagues, which is the sharpest thing on this connection. get_or_create_dm opens a conversation without sending anything, so it is the safer half of the pair |
| Spaces — Pages | Creating, reading and updating pages in a space, duplicating a page under a new title, listing a space's pages as a hierarchical tree, and reading its homepage | Read / Write | update_page replaces content colleagues work from, with no draft or suggestion step between the call and what they see. Read the page first, since an update against a stale copy overwrites whatever changed since |
| Spaces — Spaces | Creating, reading and updating spaces, listing them, and listing the users and teams collaborating on a space | Read / Write | list_space_collaborators shows who can reach a space, which is an access map worth treating as internal. Collaborators cannot be added or removed here, only listed |
Connect Teamwork.com MCP to your AI agent
Four steps, under ten minutes.
- 01
Sign in to NeuralTrust
Open the NeuralTrust platform, sign in and go to Getting started, then Connect your tools.
- 02
Choose your agent
When asked which agent you use, pick Claude Code, Codex, Cursor, Windsurf, Copilot or your own code.
- 03
Connect Teamwork.com
Find Teamwork.com in the tool list and connect it.
- 04
Connect your agent
Follow the steps on screen to add the config to your agent, then send the test prompt.
Permissions down to the single tool
You do not allow Teamwork.com MCP or block it. You allow the exact tools each agent can call, one by one.
| Teamwork.com MCP | 225 TOOLS AVAILABLE |
|---|---|
| Code review agent | 4 tools, read only |
| Issue triage bot | 9 tools |
| Platform team assistant | all 225 tools |
- Tool by toolConnecting the server is not the unit of access. Each tool is granted on its own.
- Default denyA tool you have not enabled is never exposed to the agent.
- Approval on writesTools that change or delete things can require a human to confirm.
Is Teamwork.com MCP secure?
One endpoint instead of raw credentials. Every call is checked against policy and logged.
| AI agentsConnect through your MCP client configuration. | NeuralTrustApply controls based on your configuration. | Teamwork.com MCP225 tools across 14 toolsets and 4 products, split 116 read and 109 write by Teamwork's own count. The catalogue is generated from the server's toolset registry, and Teamwork states there is no second, wider set of tools behind it. No delete tool is published at all: deletes sit behind an allowDelete flag that no shipped server enables. Six profile endpoints each serve a fixed set of toolsets on their own path, a read-only flag drops every write tool, a token carries one scope per product, and every call runs with the permissions of the authorising account. |
|---|---|---|
| Coding assistants | Connection & credentialsper user | Projects — Projects38 tools |
| Internal copilots | Tool access policiesdeny by default | Projects — Content34 tools |
| Workflow automations | Request & response inspectioninline | Projects — Tasks24 tools |
| Compatible MCP clients | Activity monitoringalerts and logs | All other toolsets129 tools |
Use Teamwork.com MCP with any AI agent framework
Setup guides for the tools your team already uses.
Claude CodeAdd Teamwork.com MCP to Claude Code from the terminal
CursorUse Teamwork.com MCP tools inside Cursor's agent
CodexConnect Teamwork.com MCP to the OpenAI Codex CLI
WindsurfGive Windsurf's Cascade access to Teamwork.com MCP
Gemini CLICall Teamwork.com MCP from Gemini CLI prompts
Google AntigravityLet Antigravity agents work with Teamwork.com MCP
GitHub CopilotUse Teamwork.com MCP in Copilot agent mode
Claude EnterpriseRoll out Teamwork.com MCP across your Claude workspace
Microsoft Copilot StudioAdd Teamwork.com MCP as a tool for Copilot Studio agents
Open WebUIBring Teamwork.com MCP tools into Open WebUI chats
LangChainLoad Teamwork.com MCP tools into LangChain agents
HaystackPlug Teamwork.com MCP into Haystack agent pipelines
Strands AgentsGive Strands agents Teamwork.com MCP tools
Vercel AI SDKCall Teamwork.com MCP from Vercel AI SDK apps
n8nUse Teamwork.com MCP in n8n AI agent workflows
Common questions about running Teamwork.com MCP through TrustGate.
Other MCP servers
More servers from the catalog.
Shortcut
Securely connect your AI agents to Shortcut through TrustGate: stories, subtasks, tasks and links, epics, iterations, objectives, labels, custom fields, teams, workflows and documents, scoped per agent and logged.
56 toolsView)
Gorgias
Securely connect your AI agents to Gorgias through TrustGate: tickets, customers, macros, helpdesk rules, help center and guidance, AI Agent config and support actions, and analytics, scoped per agent and logged.
165 toolsViewClose
Securely connect your AI agents to Close through TrustGate: leads, contacts, opportunities, calls, emails, SMS and transcripts, tasks, templates, custom objects, voice agents and billing, scoped per agent and logged.
127 toolsViewPort
Securely connect your AI agents to Port through TrustGate: blueprints and entities, self-service actions and workflows, scorecards, integrations, dashboards, audit logs and OAuth apps, scoped per agent and logged.
52 toolsView
Connect Teamwork.com MCP instantly
Pick the tools, point your agent at TrustGate, and Teamwork.com MCP is live: scoped, inspected, logged.