Todoist MCP Server

Community

Manage Todoist tasks, projects, comments, and labels via natural language with MCP

How to install ↓ Compare GitHub source repository ↗

Category: Collaboration #29 of 55
Stars: ★ 63 Popular
Transport: stdio (local process)
Runtime: Node.js
Credentials: API key / credential required
License: MIT
Last commit: 6 months ago ⚠
Tools: 36

53FMRS · D

A community-maintained MCP server based on the README, offering full Todoist REST API v2 coverage plus Sync API support across tasks, projects, sections, comments, and labels, with batch operations and natural-language interaction. Documentation is clear and multiple mainstream MCP clients are supported. Best for heavy Todoist users, with attention to its unofficial status and token security.

Strongest · Documentation 14/20 Weakest · Security and permissions 7/20

Reliability

9/20

Security and permissions

7/20

Maintenance

10/20

Documentation

14/20

Setup experience

13/20

Why each score

Reliability 9/20: The README documents a full and coherent tool surface (tasks/projects/sections/comments/labels/utils) and mentions a local dev/debug workflow via `npm run build && npx @modelcontextprotocol/inspector`, suggesting the author exercised the MCP init/tool-list handshake at least manually. However, no CI workflow files or committed automated tests are present in the supplied material, and there is no visible error-handling or dependency-failure strategy to inspect. Per the static calibration rule, reliability cannot exceed 12 without verifiable CI+test evidence, so a 9 reflects a plausible happy path with unverified edge cases and error handling.
Security and permissions 7/20: Credentials are passed via the API_KEY environment variable rather than hardcoded, which is sound practice, and no example contains a real secret. However, Todoist API tokens are inherently account-wide, and the server exposes no scoping or least-privilege options. More importantly, several irreversible destructive tools (delete_tasks, delete_projects, delete_sections, delete_comments, delete_labels, remove_shared_labels) are documented with no mention of confirmation prompts, undo, or safeguards, and there is no disclosure of network egress boundaries. These are real gaps, but nothing in the material indicates malware, covert exfiltration, or fabricated identity/results, so this does not meet the 0-4 red-line bar; it lands in the lower-mid range for 'main risks visible but confirmation/scoping incomplete.'
Maintenance 10/20: The project is MIT licensed, not archived, and published to npm with a version badge, showing basic release and license governance. However, no changelog, release history, recent commit timestamps, or maintainer responsiveness evidence is provided in the material; star count and open-issue count are discovery signals only and are not scored. Lack of verifiable ongoing-maintenance evidence caps this at a middling score.
Documentation 14/20: The README is well organized: it covers obtaining a Todoist API token, provides configuration snippets for multiple clients (Claude Desktop, Claude Code, Cursor, Codex, Gemini CLI), lists the complete tool set by category, gives example prompts, and documents a local dev workflow. It falls short on per-tool parameter/schema detail, error codes, rate limits, cost/quota disclosure, and a dedicated troubleshooting section, and leaves assumptions (token scope, Sync vs REST API differences) unexplained — hence the deduction.
Setup experience 13/20: Installation is straightforward: a single `npx -y todoist-mcp` command plus one environment variable, with explicit config examples for several mainstream clients, keeping steps and complexity low. Since this is a purely static review with no execution logs, CI runs, or integration tests to confirm the flow works end-to-end, the score is capped at 15 per the static calibration rule, landing in the 'clear steps but unverified' range.

Static review · not runListed 2026-09-07

Read the FMRS scoring method →

Fit and risk

What it can accessUses the networkChanges third-party account data

Best for

Existing Todoist users on Claude Desktop, Cursor, Codex, or Gemini CLI who want AI-driven task management
Users needing batch task operations and filter-language queries

Not for

Users who do not use Todoist
Users requiring official Todoist support or guarantees (community project, MIT licensed)

Required permissions

Requires a Todoist API token with full read/write access to tasks, projects, sections, comments, and labels in that account

Risks and side effects

API token leakage could expose task data to viewing, modification, or deletion
AI misinterpretation could accidentally delete or alter tasks and projects
Third-party community project, not maintained by Todoist

Setup

Before you start

Runtime：Node.js

API_KEY requiredsecret Your Todoist API token, found under Settings → Integrations → "Developer" in your Todoist account.

Obtain your API token from Todoist under Settings → Integrations → Developer, then add to your client's mcpServers config: command npx, args ["-y","todoist-mcp"], and env API_KEY set to your Todoist API token.

claude_desktop_config.json

{"mcpServers":{"todoist":{"command":"npx","args":["-y","todoist-mcp"],"env":{"API_KEY":"your_todoist_api_token"}}}}

Shown for Claude Desktop. Other clients may use a different file or key (VS Code uses "servers") — the configurator below converts it.

.vscode/mcp.json

{
  "servers": {
    "todoist": {
      "command": "npx",
      "args": [
        "-y",
        "todoist-mcp"
      ],
      "env": {
        "API_KEY": "your_todoist_api_token"
      }
    }
  }
}

Goes in your project's .vscode/mcp.json (VS Code uses a "servers" key).

Terminal

claude mcp add todoist -e API_KEY=your_todoist_api_token -- npx -y todoist-mcp

Run it in a terminal; replace any <…> placeholders with your own values first.

Generate and validate client config

Check that it works

Add the todoist server (npx -y todoist-mcp) with API_KEY to your client's mcpServers config and restart it; tools like get_tasks_list and create_tasks should appear in the tool list, and asking "What projects do I have in my Todoist?" should return your projects.

Troubleshooting

Verify the API_KEY environment variable is set to a valid Todoist API token
Check the token in Todoist under Settings → Integrations → Developer
Ensure Node.js is installed and npx is available
Use @modelcontextprotocol/inspector to debug server connectivity

Things to try

Once connected, you can ask your AI assistant things like:

What tasks do I have due today?
Create a task to review the quarterly report by next Friday
Make a new project called 'Home Renovation'
Mark my dentist appointment task as complete

Tools 36

get_tasks_list read-only

Get tasks with optional filtering by project, section, label, or IDs

get_tasks_by_filter read-only

Query tasks using Todoist filter language (e.g. today, overdue, P1 | P2)

create_tasks writes

Create new tasks with various attributes

get_tasks read-only

Get specific tasks by ID or name

update_tasks writes

Update existing tasks

close_tasks writes

Mark tasks as complete

reopen_tasks writes

Reopen completed tasks

delete_tasks destructive

Delete tasks

Show 28 more tools

move_tasks writes

Move tasks to a different project or section

get_projects_list read-only

Get all projects

create_projects writes

Create new projects

get_projects read-only

Get specific projects by ID or name

update_projects writes

Update existing projects

delete_projects destructive

Delete projects

get_collaborators read-only

Get all collaborators for a project

move_projects writes

Move projects to a different location or subproject

get_sections_list read-only

Get all sections or filter by project

create_sections writes

Create new sections

get_sections read-only

Get specific sections by ID or name

update_sections writes

Update sections

delete_sections destructive

Delete sections

get_comments_list read-only

Get comments for a project or task

create_comments writes

Create new comments

get_comments read-only

Get specific comments by ID

update_comments writes

Update comments

delete_comments destructive

Delete comments

get_labels_list read-only

Get all personal labels

create_labels writes

Create new personal labels

get_labels read-only

Get personal labels by ID or name

update_labels writes

Update personal labels

delete_labels destructive

Delete personal labels

get_shared_labels read-only

Get all shared labels

rename_shared_labels writes

Rename shared labels

remove_shared_labels destructive

Remove shared labels

utils_get_colors read-only

Get available colors for projects, labels, and filters

projects_list read-only

Prompt returning projects with their sections and params in markdown format

Use cases

Query today's or high-priority tasks conversationally

Create, complete, or reopen tasks through natural language

Manage projects, sections, comments, and labels

Batch-retrieve tasks using Todoist filter language

Supported clients

Claude Desktop

Claude Code

Cursor

Codex

Gemini CLI

Listed from the project's documentation, not tested by this site.

Overview

A community-maintained Model Context Protocol (MCP) server that integrates Claude and other AI assistants with Todoist. It provides a full implementation of the Todoist REST API v2 and supports the Todoist Sync API, enabling natural-language creation, update, closure, reopening, moving, and deletion of tasks, plus management of projects, sections, comments, and personal/shared labels. On startup it supplies your projects and labels to the AI via server instructions, and it supports batch processing and search by name instead of ID.

Similar servers

Atlassian Rovo MCP Server 66 · C

Securely connect Atlassian Cloud to AI tools.

★ 1.1k Compare with this →

Notion MCP Server 64 · C

Official Notion MCP server for querying, editing, and managing Notion pages and databases via AI agents.

★ 4.7k · Tools 9 Compare with this →

WhatsApp Channel for Claude Code 64 · C

Drive your Claude Code session straight from WhatsApp

★ 109 · Tools 10 Compare with this →

Anytype MCP Server 64 · C

Official MCP server that lets AI assistants manage your encrypted Anytype wiki in natural language.

★ 533 Compare with this →

Compare with other MCP servers

Source revision 3c113e810c43 Data synced 2026-10-19 Read the FMRS scoring method