Skip to main content
The Model Context Protocol (MCP) is an open standard that lets AI tools interact with external services. Trilo’s MCP server gives your AI coding assistant direct access to your workspace — create tasks, read pages, send messages, and more without leaving your editor.

What You Can Do

Once connected, your AI tool can interact with your entire Trilo workspace:
CategoryExample Prompts
Tasks”Create a bug report for the login issue” / “What tasks are assigned to me?”
Pages”Write an implementation plan and save it as a page” / “Update the API docs page”
Messages”Send a status update to the backend channel” / “What did the team discuss today?”
Calendar”Schedule a design review for Friday at 2pm” / “What meetings do I have this week?”
Meetings”Summarize the standup transcript” / “Search meetings for discussion about auth”
Spreadsheets”Create a sprint tracking spreadsheet” / “Update the budget numbers”
Members”Who’s on the mobile project?” / “Add Sarah to the backend project”
Knowledge Base”Search our docs for the deployment guide” / “Find resources about onboarding”

Supported AI Clients

Claude Code

Anthropic’s CLI tool. Supports PAT authentication via claude mcp add.

Claude Desktop

Configure via JSON config file with PAT authentication.

Cursor

Configure via Settings UI or .cursor/mcp.json with PAT authentication.

Windsurf

Configure via Settings or config file with PAT authentication.
Any MCP-compatible client can connect using the HTTP endpoint with a Personal Access Token.

Authentication

Trilo uses Personal Access Tokens (PATs) to authenticate MCP connections. PATs give you fine-grained control over permissions and can be revoked at any time.
MethodBest ForSetup
Personal Access Token (PAT)All AI coding tools, CI/CD, scriptsGenerate in Trilo settings, pass as Bearer token

Token Scopes

When creating a PAT, you choose which scopes to grant:
ScopePermissions
mcp:readList and read workspaces, projects, tasks, pages, messages, members, events, meetings, resources, spreadsheets
mcp:writeEverything in mcp:read plus create and update tasks, pages, messages, events, spreadsheets, and project members
mcp:adminEverything in mcp:write plus administrative operations
For most development workflows, mcp:read and mcp:write together provide full access to all tools. Use mcp:read alone if you only want your AI tool to read data without making changes.

Next Steps