Todoist MCP Server
This server acts as an AI-agent-friendly Model Context Protocol (MCP) interface for Todoist, providing comprehensive task management with fine-grained permission control and optimized data payloads.
Task Management
List, retrieve, create, update, complete, reopen, delete, move, and search tasks
Supports due dates, priorities (P1–P4), labels, descriptions, durations, subtasks, and natural language quick-add
Batch create multiple tasks at once; search tasks scoped to a project, section, or label
Project Management
List, get, create, update, and delete projects
Configure name, color, parent, view style (list/board), and favorite status
Section Management
List, get, create, update, and delete sections within projects
Comment Management
List, get, create, update, and delete comments on tasks or projects
Supports optional prefix tags (e.g.,
[Research],[Context],[Note]) and file attachment metadataShortcut helpers for adding research notes and context information to tasks
Label Management
List, create, update, and delete personal labels with name, color, order, and favorite status
Completed Tasks & Statistics
List completed tasks with filters (date range, project, section, pagination)
Retrieve productivity statistics related to completed tasks
Advanced Todoist Surfaces
Read access to reminders, saved filters, activity logs, backups, and user/plan information
Tool Control & Performance
Server-side permission model (
read,write,destructivetiers) controls which tools are registeredConfigurable via mode (
full,standard,read-only) or explicit allow/deny listsDefaults to slim, token-efficient JSON payloads; raw payloads available on request
Audit the live tool surface via
todoist_get_capabilities
Provides integration with the Todoist REST API v2, allowing for the management of tasks, projects, sections, comments, and labels, as well as performing batch operations and accessing completed task history.
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Todoist MCP ServerAdd 'Review the budget proposal' to my Work project for tomorrow morning"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Todoist MCP Server (v2)
A Model Context Protocol (MCP) server for the unified Todoist API v1, built for AI-agent consumption: a declarative tool registry with a server-side permission model, slim token-efficient payloads, and coverage of the Pro surfaces (reminders, saved filters, activity log, backups, productivity stats, natural-language quick-add).
v2 highlights (2026-07-05 rebuild)
Permission model at the registration layer. Every tool carries a tier (
read/write/destructive); tools outside the configured mode are never registered, which holds even on harness channels that skip client-side permission prompts.todoist_get_capabilitiesaudits the live surface (registered + filtered + why).Slim payloads by default. Todoist objects carry 25+ fields; slim mode returns only decision-relevant ones and uses compact JSON. Example: 4 projects = ~770B vs ~6KB raw pretty-printed.
TODOIST_SLIM=0restores raw passthrough.Reliable comment counts. The v1 list-payload
note_countis unreliable (reports 0 on tasks with comments); slim list payloads drop it, andtodoist_get_taskreturns an authoritativecomment_countfetched from/comments.Fixed v1.x defects.
todoist_search_tasksnow honorsproject_id/section_id/label(previously accepted-and-ignored);todoist_get_completed_statspages up to 1000 completions (previously silently capped at 200).New v1 surfaces. Reminders, saved filters (Sync-API lane — filters have no REST resource, verified live), activity log (
/activities), backups, user/plan info, productivity stats (/tasks/completed/stats),todoist_quick_add(natural language), anddeadline_dateon task create/update (the do-date vs drop-dead-date distinction).
Related MCP server: Enhanced Todoist MCP Server Extended
Configuration
{
"mcpServers": {
"todoist": {
"command": "node",
"args": ["/path/to/todoist-mcp/dist/index.js"],
"env": { "TODOIST_API_TOKEN": "${TODOIST_API_TOKEN}" }
}
}
}Environment knobs (all optional):
Var | Values | Default | Effect |
|
|
|
|
| comma list of tool names | — | Force specific tools OFF (wins over everything). |
| comma list of tool names | — | Force specific tools ON regardless of mode (deny still wins). |
|
|
| Slim vs raw payloads. |
Never commit your API token.
Tool surface
51 tools registered in standard mode; 9 more (delete_*, reopen_*) in full.
Ask the server itself: call todoist_get_capabilities for the live list with tiers.
Groups: tasks (list/get/search/hierarchy/duplicates · create/update/quick-add/complete/move/
follow-up/with-context · 4 batch ops · deletes+reopens destructive-tier), projects, sections,
comments (incl. [Research]/[Context] helpers), labels, completed+stats, workspace overview,
reminders, saved filters, activity log, backups, user.
Priority convention: API values are inverted from the UI (P1 urgent = 4 … P4 untriaged = 1).
Response format
Every tool returns { success, data } or { success: false, error: { code, message, details? } }
as a single compact-JSON text block.
Development & testing
npm run build # tsc
npm run typecheck
node smoke-test.mjs # protocol-level live smoke: spawns dist/index.js over stdio,
# verifies mode filtering + live reads (needs TODOIST_API_TOKEN)License
ISC
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Latest Blog Posts
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/DevonGroff/todoist-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server