otheroauth2

Todoist MCP

Todoist MCP lets agents read, create, update, organize, and complete tasks and projects through Todoist's hosted Streamable HTTP server.

Verdict

The Todoist MCP connects your team's task manager to Switchy, letting AI read, create, and update tasks across projects without leaving the conversation. @mention it to pull overdue items, add tasks with due dates and labels, or reorganize priorities based on chat context. Most useful for teams already running their work through Todoist who want AI to handle routine task hygiene — checking what's blocked, surfacing what's due today, or capturing action items from meeting notes. Requires OAuth with full task read/write access, so review scope carefully if your Todoist holds sensitive client work.

Common use cases

  • Capture action items from meeting transcripts
  • Surface overdue tasks at daily standup
  • Reprioritize sprint backlog based on discussion
  • Add labeled tasks from customer feedback
  • Check what's due today across all projects

Integration

Vendor
Todoist MCP
Category
other
Auth
OAUTH2
Composio slug
todoist_mcp

Tools

Per-tool listings haven't synced yet for Todoist MCP. The connection itself works - your Project can already @-mention it. Tool descriptions will fill in on the next Composio ingest.

Setup

Setup guide

  1. 11. Open your Switchy workspace and navigate to Settings > Integrations. 2. Find Todoist MCP in the list and click Connect. 3. You'll redirect to Todoist's OAuth consent screen — sign in with the account that owns your team's projects. 4. Grant the requested scopes: read tasks, write tasks, and read projects (the MCP needs all three to function). 5. After authorizing, Switchy redirects you back and shows a green Connected badge. 6. Open any Project in Switchy and type '@Todoist show my overdue tasks' in the chat. 7. If the MCP returns a list, the connection works. 8. To invoke it going forward, @mention Todoist in any message and describe the task operation you want — the AI will route your request to the appropriate tool.

What teammates see: by default, memories from Todoist MCP are scoped to the Project (PROJECT visibility) - you can mark any memory PRIVATE or share it ORG-wide.

Works well with

Top models

Compatibility data appears once enough Projects have used this MCP together with a given model.

How Switchy teams use it

Not enough Projects yet to publish anonymised usage stats (we require ≥ 50 Projects per week).

Starter prompts

Triage Overdue Work

@Todoist show me everything overdue across all projects. For each task, tell me the project, priority, and how many days it's been overdue.

Capture Meeting Actions

@Todoist create tasks for these action items: finalize Q2 budget by Friday, schedule client demo next Tuesday, review contract edits by end of week. Assign all to the Operations project.

Daily Standup Summary

@Todoist list all tasks due today, grouped by project. Include priority levels and any labels.

Reprioritize Sprint Backlog

@Todoist find all tasks in the Sprint project labeled 'backend' and set their priority to P1. Then show me the updated list sorted by due date.

Add Customer Feedback Tasks

@Todoist create a task in the Product Feedback project: 'User requests dark mode for mobile app'. Label it 'feature-request' and set due date to next Monday.

Example outputs

Illustrative - representative of the model's voice and quality, not literal recordings.

Prompt

@todoist show me all my overdue tasks grouped by project

Output

Here's what I found across your Todoist account: **Work** (3 overdue) - Finish Q4 budget review (due 2 days ago) - Send client proposal draft (due yesterday) - Update team wiki (due 4 days ago) **Personal** (1 overdue) - Schedule dentist appointment (due 5 days ago) All tasks are currently assigned to you. Would you like me to reschedule any of these or mark priorities?

Notes

This read-only query demonstrates cross-project visibility. Requires OAuth scope for reading tasks and projects. Response reflects your actual task state at query time — overdue counts and project names will vary by account.

Prompt

@todoist create a task 'Review PR #847' in my Engineering project, due tomorrow at 2pm, priority 2

Output

Task created successfully in your Engineering project: **Review PR #847** - Due: Tomorrow at 2:00 PM - Priority: P2 (high) - Project: Engineering - Task ID: 7234891023 The task is now visible in your Todoist inbox and will trigger a notification 30 minutes before the due time based on your account settings.

Notes

This write operation requires OAuth scope for task creation. The MCP respects your Todoist notification preferences but cannot override them. Priority levels (1-4) map to Todoist's P1-P4 system where P1 is urgent.

Prompt

@todoist look at my tasks due this week and draft a Friday status update email summarizing what I completed versus what's still open

Output

Based on your Todoist activity this week: **Completed (7 tasks)** - Shipped user authentication refactor - Reviewed 3 design mockups - Conducted two customer interviews - Finalized hiring panel feedback **Still Open (4 tasks)** - Deploy staging environment (due Friday) - Write API documentation (due Thursday) - Two code reviews pending Draft email: *Subject: Week of Jan 13 – Status Update* This week I completed the authentication refactor and wrapped customer research for the Q1 roadmap. Still working through API docs and staging deployment before EOW. Code reviews in progress. Would you like me to adjust the tone or add detail to any section?

Notes

This synthesis example combines task data with AI reasoning to generate narrative summaries. The MCP provides raw task metadata (titles, completion status, due dates); the AI interprets context and drafts prose. Accuracy depends on how you title tasks in Todoist.

Use-case deep-dives

Async standup for remote teams

When Todoist MCP fits distributed daily check-ins

A 6-person engineering team across three time zones posts daily updates in Slack, referencing tasks they closed or blocked. The Todoist MCP lets the team's shared AI pull task status, due dates, and assignees without switching apps. This works when your standup ritual is already text-based and your task metadata lives in Todoist. The OAuth2 flow means each person authenticates once, then the AI sees their personal projects and shared team boards. The trade-off: if your team uses Linear or Jira for sprint tracking, you're duplicating effort by keeping Todoist in sync. Use this when Todoist is already your single source of truth for who-owns-what, and you want the AI to surface blockers or overdue items during the standup thread.

Client deliverable tracking for agencies

How agencies use Todoist MCP for milestone visibility

A 4-person creative agency juggles 8 client projects, each with its own Todoist board and recurring deliverable checklist. The MCP lets the team ask the AI 'what's due for Client X this week' or 'show overdue tasks across all projects' without opening Todoist. This scenario wins when your agency already lives in Todoist and your clients don't need direct access to the tool. The OAuth2 requirement means the AI can filter by assignee or project scope, so the account lead sees only their clients. The boundary: if you need time tracking or invoice generation tied to tasks, Todoist alone won't close the loop. Use this when your workflow is task-completion-driven and you want the AI to triage what's at risk across multiple boards.

Personal productivity coaching in Slack

When Todoist MCP supports individual habit tracking

A solopreneur or small team lead uses Todoist for personal goals (exercise, writing, admin tasks) and wants the AI to nudge them or summarize progress. The MCP connects the AI to their Todoist account so they can ask 'what did I finish this week' or 'remind me of my top 3 priorities' in Slack or the Switchy workspace. This works when the person already maintains Todoist as their personal system and wants conversational access without opening the app. The OAuth2 flow keeps their tasks private. The limit: if the team needs shared visibility into these habits (like a manager checking direct-report goals), Todoist's collaboration features are light. Use this when the individual is disciplined about logging tasks and wants the AI to reflect their own data back to them.

Frequently asked

What can the Todoist MCP do in Switchy?

The Todoist MCP connects your Todoist account to Switchy's AI workspace, letting Claude read your tasks, projects, and labels. Your team can ask questions about deadlines, filter tasks by project or priority, and get summaries of what's overdue or coming up. It doesn't create or modify tasks — it's read-only access for now.

Does the Todoist MCP need admin access to my workspace?

No. The OAuth flow asks for read-only permissions on your personal Todoist account. If you're using Todoist Business, each team member connects their own account — there's no workspace-level admin requirement. The MCP only sees tasks and projects visible to the account that authenticated.

Can the Todoist MCP create or complete tasks?

Not yet. The current integration is read-only, so Claude can surface task data but can't mark items complete, add new tasks, or change due dates. If you need write access, use Todoist's API directly or wait for Switchy to expand the tool set in a future release.

Why use the Todoist MCP instead of just opening Todoist?

The MCP lets Claude answer questions across your task list without you manually filtering or searching. Ask "what's blocking the launch" or "show overdue tasks for the design project" and get instant answers in chat. It's faster than switching tabs, especially when you're already working in Switchy with other context loaded.

Who on my team should connect the Todoist MCP?

Anyone who manages tasks in Todoist and wants Claude to reference them during planning or standups. Each connection is per-user, so if three people connect their accounts, Claude can only see their individual task lists — not a shared team view unless you all share the same Todoist projects.

More other MCPs

See all other MCPs →

Data last verified 62 hours ago.Sources aggregated hourly to weekly. See docs/architecture/model-directory.md.