TaskPile
Pile in. Sort out. Taskpile is an AI-first productivity task manager that lives inside your chat client — capture in Claude, plan in ChatGPT, finish in the app. 57 tools cover task CRUD, projects, tags, teams, bulk ops, and a morning digest.
README
Taskpile MCP
Public documentation for Taskpile's Model Context Protocol server. Taskpile is an AI-first task manager designed to be used through your chat client — capture tasks in Claude, plan in ChatGPT, finish in Taskpile.
- App: https://taskpile.app
- MCP endpoint:
https://taskpile.app/api/mcp - Transport: Streamable HTTP (JSON-RPC 2.0)
- Protocol versions supported:
2025-06-18,2025-03-26,2024-11-05
This repository contains only public-facing documentation for the MCP integration. The Taskpile app itself is a hosted commercial product — its source isn't here, but everything you need to connect a client and use the MCP tools is.
Connecting
Any MCP-compatible client can connect by pointing at the URL above. Authentication uses OAuth 2.1 with Dynamic Client Registration + PKCE — the client handles this automatically; users sign in to Taskpile in a browser tab when prompted.
Verified clients:
| Client | Status |
|---|---|
| Claude.ai (Connectors) | ✅ |
Claude Desktop (via mcp-remote) |
✅ |
| Claude Code | ✅ |
| ChatGPT (Connectors) | ✅ |
| Le Chat (Mistral) | ✅ |
| Cursor | ✅ |
| Zed | ✅ |
| MCP Inspector | ✅ |
| Lovable.app | ✅ |
Quick start by client
See examples/ for ready-to-paste config snippets:
For hosted clients like Claude.ai and ChatGPT, add Taskpile via their built-in connector UI and paste the URL above; no local config needed.
Tools
Taskpile exposes 57 tools spanning task CRUD, projects, tags, search, bulk operations, teams, and account integrations like morning digest. The full machine-readable schema (names, descriptions, JSON Schema for inputs) is in tools.json — auto-generated from the live server, so it's always in sync with what tools/list returns.
A quick taste:
create_task,list_tasks,update_task,complete_task,delete_taskassign_horizon(today_morning, tomorrow, next_week, etc.),delegate_task,set_recurrencelist_projects,create_project,archive_projectsearch_tasks(free-text + status/horizon filters)bulk_complete,bulk_delete,bulk_assign_horizon(up to 200 ids per call)changes_since(incremental sync),get_today,get_review_queueget_morning_digest,configure_morning_digestwhoami,get_inbox_email- Teams:
create_team,invite_to_team,accept_invitation,assign_task,convert_project_to_shared, etc. searchandfetch(ChatGPT Connectors built-in tools, mapped to Taskpile)
Conventions
A few project-wide conventions worth knowing if you're writing a prompt:
- Tags use
@tag_name. Projects use#project_name. Do NOT use#for tags —#foois silently treated as a project name. create_taskparses#projectand@tagshorthand from the title. Only the FIRST#token becomes the project; the rest are stripped.- To set tags at creation, pass
tags: string[]and/or put@tagin the title. Both forms work and are de-duplicated. update_taskdoes not parse#/@shorthand from the title. Usetags/addTags/removeTagsandprojectIdto change those.- Prefer dedicated verbs over
update_taskwhen they exist:complete_task,assign_horizon,delegate_task. - Bulk endpoints accept up to 200 ids per call.
Auth flow
If you're building a client and want to know what to expect:
- Client sends an unauthenticated request to
https://taskpile.app/api/mcpand receives401withWWW-Authenticate: Bearer realm="taskpile", resource_metadata="https://taskpile.app/.well-known/oauth-protected-resource". - Client fetches
https://taskpile.app/.well-known/oauth-protected-resourceand…/oauth-authorization-serverto discover endpoints. - Client POSTs to
https://taskpile.app/oauth/registerfor Dynamic Client Registration (RFC 7591). - Client kicks off the standard authorization-code + PKCE flow, opens a browser at
/oauth/authorize, exchanges the code at/oauth/token, and receives anaccess_token. - Subsequent requests carry
Authorization: Bearer <token>. Tokens are valid for 30 days; no refresh-token flow — clients re-run the authorize/token dance on expiry.
HEAD /api/mcp returns 200 without auth and is intended for liveness probes only.
Privacy + data handling
- Taskpile is GDPR-compliant. Tasks, projects, and tags created via MCP are user-owned and visible only to the authenticated user (and their team members for shared projects).
- The MCP integration sends task content to Taskpile's servers (EU-hosted). It does NOT route through Anthropic, OpenAI, or any other LLM provider — those are clients of the MCP server, not the server itself.
- Personal access tokens can be created in Taskpile → Settings → Integrations and revoked at any time.
Issues and feedback
This repo is for documentation. Bug reports about the MCP server, feature requests for new tools, or questions about the API — please open an issue here and we'll triage.
For Taskpile app issues (unrelated to MCP), please contact us through the app at https://taskpile.app.
License
Documentation in this repository is MIT-licensed (see LICENSE). The Taskpile app itself is a hosted commercial product and is not licensed for redistribution.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.