M
MCP Directory CURATED DIRECTORY
911 MCP servers indexed
中文
← Back to directory
T

Todoist MCP Server

Community
Manage Todoist tasks and projects through natural language with Claude
How to install ↓ Compare GitHub source repository ↗
Category
Collaboration #8 of 55
Stars
★ 246 Popular
Transport
stdio (local process)
Runtime
Node.js
Credentials
API key / credential required
License
MIT
Last commit
⚠
Tools
19
61FMRS · C

A community-maintained Todoist MCP server offering 19 tools spanning tasks, projects, and collaboration, with practical features like bulk operations and a Dry-Run simulation mode. The maintainer enforces a strict contribution process (single-concern PRs, verification against the live API), and documentation is thorough, but users are responsible for managing the API token securely and for the data-change risk of bulk operations.

Strongest · Documentation 16/20 Weakest · Reliability 9/20

Reliability
9/20
Security and permissions
10/20
Maintenance
13/20
Documentation
16/20
Setup experience
13/20
Why each score
Reliability 9/20
The README describes 19 MCP tools with grouped actions and a dry-run mode that simulates mutations while validating against real data, which is a good design for diagnosability. However, no server manifest was provided, no actual source code (e.g., the handshake logic in src/index.ts) was shown, and no CI workflow or test file contents were visible — only README mentions of `npm test` commands. Per the static calibration rule, without verifiable execution evidence (real CI plus committed tests covering key paths), reliability is capped at 12; given the complete inability to verify that declared tools match real behavior, it is scored 9.
Security and permissions 10/20
Credentials are passed via the `TODOIST_API_TOKEN` environment variable, and examples use placeholders rather than real secrets — no credential leakage or hardcoding observed. Dry-run mode is a positive safety feature enabling side-effect-free validation. However, dry-run is off by default, and irreversible operations such as `todoist_task_bulk` (bulk_delete) and `todoist_task` (delete) show no documented additional confirmation step in default mode; the README also does not disclose outbound data flow (e.g., telemetry/logging). Since task/project deletion affects only personal data with limited blast radius, this does not rise to red-line severity (0-4), but evidence of least-privilege design and confirmation for dangerous operations is insufficient, so it is scored 10.
Maintenance 13/20
The project is MIT-licensed, has 244 stars and 19 open issues, is not archived, and includes a CHANGELOG.md plus a detailed contribution guide requiring tests, passing CI, and single-concern PRs — signs of active governance. However, no evidence of actual release cadence, recent commit activity, or a dedicated security-response channel (e.g., SECURITY.md) was provided, and dependency-update frequency cannot be confirmed from the given material, so it is scored 13 rather than higher.
Documentation 16/20
The README is thorough, covering installation (one-click install, manual JSON config, both major platforms), authentication (token acquisition link), a detailed dry-run mode explanation, a full tool overview table for all 19 tools, extensive natural-language usage examples, a troubleshooting section, and pointers to layered developer docs (TOOLS_REFERENCE.md, architecture breakdown). Rate limits, cost/quota considerations, and security boundaries are disclosed only sparingly, and detailed tool parameter constraints require consulting an external file not verifiable in this review, so it does not receive full marks and is scored 16.
Setup experience 13/20
The install path is clear: one-click .mcpb install for Claude Desktop, a `claude mcp add` CLI command, or manual JSON configuration, with config file paths listed for major platforms — a low barrier to entry. However, per the static calibration rule, without verifiable CI/test execution evidence, setup is capped at 15; combined with the lack of a manifest to confirm the tools actually register as documented, this uncertainty brings the score down to 13.

Static review · not runListed 2026-08-26

Read the FMRS scoring method →

Fit and risk

What it can accessUses the networkChanges third-party account data

Best for

  • Existing Todoist users or teams who want to manage tasks and projects through conversation with Claude
  • Users who need to run bulk task operations (e.g., mass priority updates, clearing overdue tasks)
  • Developers who want to validate automations in Dry-Run mode before touching real data

Not for

  • Users who don't use Todoist, or who need to connect to other task tools like Jira or Asana
  • Enterprise scenarios requiring an official Todoist product or SLA-backed support
  • Highly privacy-sensitive setups that don't want API tokens and task data routed through a third-party MCP server

Required permissions

  • Requires a personal Todoist API token (from Todoist developer settings), which grants account-level read/write access
  • Can read and modify all account data including tasks, projects, sections, labels, comments, reminders, filters, notes, and shared labels
  • Some features (reminders, filters, shared labels) require a Todoist Pro or Business subscription

Risks and side effects

  • This is a third-party open-source project maintained by an individual developer, not an official Todoist product
  • The API token is stored in plaintext in the client config file and should be protected from exposure
  • Bulk delete or bulk complete operations can make irreversible changes to real Todoist data if misconfigured; test with DRYRUN=true first
  • Task and comment content passes through Claude, so users with sensitive data should evaluate their own data-flow risk

Setup

Before you start

Runtime：Node.js

TODOIST_API_TOKEN requiredsecret Required Todoist developer API token, obtained from Todoist Settings > Integrations > Developer.
Other optional settings (1)
DRYRUN optional Set to true to enable dry-run mode where mutations are simulated; optional.

1) Claude Desktop users can download the todoist-mcp.mcpb file from the latest GitHub release and double-click or drag it onto Claude Desktop, then enter a Todoist API token when prompted; 2) Claude Code or other MCP clients can run claude mcp add todoist -e TODOIST_API_TOKEN=your_token -- npx @greirson/mcp-todoist; 3) alternatively, edit claude_desktop_config.json or ~/.claude.json manually, adding an mcpServers entry with command npx, args ["@greirson/mcp-todoist"], and env TODOIST_API_TOKEN set to your token; 4) to test safely without affecting real data, add DRYRUN=true to the env block to enable dry-run mode.

claude_desktop_config.json 
{"mcpServers":{"todoist":{"command":"npx","args":["@greirson/mcp-todoist"],"env":{"TODOIST_API_TOKEN":"your_api_token_here"}}}}

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": [
        "@greirson/mcp-todoist"
      ],
      "env": {
        "TODOIST_API_TOKEN": "your_api_token_here"
      }
    }
  }
}

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

Terminal
claude mcp add todoist -e TODOIST_API_TOKEN=your_api_token_here -- npx @greirson/mcp-todoist

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

Generate and validate client config

Check that it works

After configuring, ask your client "Show me my Todoist projects"; a real project list proves the TODOIST_API_TOKEN works and the server is connected. You can also run the todoist_utility test_connection action.

Troubleshooting

  1. If you see "No Todoist projects found" or connection errors, verify the API token is correct and free of extra spaces or quotes in the config
  2. If the MCP server doesn't load, confirm the package is installed globally (npm list -g @greirson/mcp-todoist) and fully restart Claude Desktop
  3. Check that the config file path matches your OS (macOS/Windows/Claude Code differ), and try the full path to the mcp-todoist binary if needed
  4. On macOS/Linux, permission errors may require manually creating the config directory (e.g., mkdir -p ~/.config) and ensuring Claude Desktop can read it

Things to try

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

  • Show me all my Todoist projects
  • Create task 'Team Meeting' in project 12345 due tomorrow with label 'Urgent'
  • Quick add: Review PR next Monday #Work @code-review p2 //Check error handling
  • Complete all tasks containing 'review' in project 12345

Tools 19

todoist_task writes
Create, get, update, delete, complete, and reopen tasks, with priorities, due dates, labels, and Quick Add natural-language parsing
todoist_task_bulk destructive
Bulk create, update, delete, and complete tasks for efficient multi-task operations
todoist_subtask writes
Create subtasks, bulk-create, convert tasks to subtasks, promote subtasks to tasks, and view task hierarchy
todoist_project destructive
Project create, get, update, delete, archive, and collaborator management
todoist_project_ops writes
Advanced project operations: reorder, move to parent, get archived projects
todoist_section destructive
Section create, get, update, delete, move, reorder, and archive
todoist_label destructive
Label create, get, update, delete, and usage statistics
todoist_comment destructive
Create, get, update, and delete task/project comments
Show 11 more tools
todoist_reminder destructive
Reminder (Pro) create, get, update, and delete
todoist_filter destructive
Custom filter (Pro) create, get, update, and delete
todoist_collaboration writes
Team collaboration features: invitations, notifications, and workspace operations
todoist_user read-only
User profile info, productivity stats, and karma history
todoist_utility read-only
Connection, feature, and performance testing, plus duplicate task find/merge utilities
todoist_activity read-only
Retrieve activity log, events, and activity summaries for audit trails
todoist_task_ops writes
Advanced task operations: move, reorder, close
todoist_completed read-only
Retrieve completed tasks, get all completed tasks, and completion statistics
todoist_backup read-only
List and download Todoist's automatic backups
todoist_notes destructive
Create, get, update, and delete project notes (for collaborators)
todoist_shared_labels destructive
Workspace shared label (Business) create, get, rename, and remove

Use cases

Create, update, and complete Todoist tasks and subtasks using natural language
Bulk create, update, or delete multiple tasks for efficiency
Quickly capture tasks with dates, projects, labels, and priority via Quick Add syntax
Manage projects, sections, labels, comments, and reminders
Validate bulk operations and automation scripts with Dry-Run mode before running them for real

Supported clients

Claude Desktop
Claude Code

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

Overview

An open-source MCP server that connects Claude to Todoist for complete task and project management through natural language. It exposes 19 tools covering tasks, subtasks, projects, sections, labels, comments, reminders, filters, collaboration, user profile, activity logs, completed tasks, backups, notes, and shared labels, with support for bulk operations and Todoist's Quick Add natural-language parsing. A built-in Dry-Run mode lets you validate operations against real data without making actual changes.

Similar servers

Compare with other MCP servers

Source revision fc8d80599334 Data synced 2026-10-19 Read the FMRS scoring method