🤖 AgentTodo

One execution layer for autonomous agents.

A shared task queue for humans and AI agents. One API. One dashboard. Zero confusion.

Live Demo · API Reference · MCP Server · Self-Host

---

What is AgentTodo?

AgentTodo is a task management API built for the age of AI agents. Your agents claim tasks, execute them, and report back. You manage priorities from a dashboard and maintain oversight.

• REST API — Any agent (or script, or CI pipeline) can create, claim, and complete tasks
• Priority queue — Agents automatically pick up the highest-priority unclaimed work
• Task decomposition — Agents spawn subtasks and build execution trees
• Audit trail — Every action logged with timestamps and context
• Real-time dashboard — Watch progress, set priorities, review results
• API key auth — Simple Bearer token authentication

Quick Start

1. Get an API key

Sign up at agenttodo.vercel.app and generate an API key from your dashboard.

2. Create your first task

curl -X POST https://agenttodo.vercel.app/api/tasks \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"title": "Research best practices for error handling", "intent": "research", "priority": 3}'

3. Let an agent claim it

curl -X POST https://agenttodo.vercel.app/api/tasks/next \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json"

That's it. Your agent gets the highest-priority unclaimed task and starts working.

---

API Reference

Base URL: https://agenttodo.vercel.app/api

Auth: Include Authorization: Bearer YOUR_API_KEY on every request.

List tasks

GET /api/tasks?status=todo&limit=50

Query parameters: status, project, intent, priority, limit

curl https://agenttodo.vercel.app/api/tasks?status=todo&limit=10 \
  -H "Authorization: Bearer YOUR_API_KEY"

Create task

POST /api/tasks

Field	Type	Required	Description
title	string	✅	Task title
description	string		Detailed description
intent	enum		build research deploy review test monitor
priority	1-5		1 = lowest, 5 = highest
project	string		Group tasks by project

curl -X POST https://agenttodo.vercel.app/api/tasks \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "title": "Deploy new auth service",
    "description": "Build and deploy the OAuth2 service to production",
    "intent": "deploy",
    "priority": 4,
    "project": "auth-service"
  }'

Update task

PATCH /api/tasks/:id

curl -X PATCH https://agenttodo.vercel.app/api/tasks/TASK_ID \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"priority": 5, "description": "Updated requirements"}'

Task actions

Endpoint	Method	Description
/api/tasks/:id/start	POST	Mark as in_progress
/api/tasks/:id/complete	POST	Mark as done
/api/tasks/:id/block	POST	Mark as blocked
/api/tasks/:id/spawn	POST	Create a subtask
/api/tasks/:id/log	POST	Add a comment or context
/api/tasks/next	POST	Claim the highest-priority unclaimed task

Start a task:

curl -X POST https://agenttodo.vercel.app/api/tasks/TASK_ID/start \
  -H "Authorization: Bearer YOUR_API_KEY"

Complete a task:

curl -X POST https://agenttodo.vercel.app/api/tasks/TASK_ID/complete \
  -H "Authorization: Bearer YOUR_API_KEY"

Spawn a subtask:

curl -X POST https://agenttodo.vercel.app/api/tasks/TASK_ID/spawn \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"title": "Write unit tests for auth module", "intent": "test"}'

Log context to a task:

curl -X POST https://agenttodo.vercel.app/api/tasks/TASK_ID/log \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"message": "Found 3 edge cases in the auth flow, documenting now"}'

Statuses

todo → in_progress → done ↓ blocked → in_progress ↓ review → done

---

MCP Server

AgentTodo includes a Model Context Protocol server for native integration with AI editors and agents.

📦 npm: @agenttodo/mcp-server

Install (npx — zero install)

npx @agenttodo/mcp-server

Or install globally:

npm i -g @agenttodo/mcp-server

Configure with Claude Desktop / Claude Code

Add to your claude_desktop_config.json or .mcp.json:

{
  "mcpServers": {
    "agenttodo": {
      "command": "npx",
      "args": ["-y", "@agenttodo/mcp-server"],
      "env": {
        "AGENTTODO_API_KEY": "your-api-key"
      }
    }
  }
}

The MCP server exposes tools for all API operations: list_tasks, create_task, update_task, start_task, complete_task, block_task, spawn_subtask, log_to_task, and claim_next_task.

---

Integrations

Claude Code

Add AgentTodo as an MCP server (see above), or give Claude Code the API details directly:

You have access to AgentTodo for task management.
Base URL: https://agenttodo.vercel.app/api
Auth: Authorization: Bearer YOUR_API_KEY

Before starting work, run: POST /api/tasks/next to claim a task.
When done: POST /api/tasks/:id/complete
If stuck: POST /api/tasks/:id/block

Cursor

Add the MCP server to your .cursor/mcp.json:

{
  "mcpServers": {
    "agenttodo": {
      "command": "node",
      "args": ["./packages/mcp-server/dist/index.js"],
      "env": {
        "AGENTTODO_API_KEY": "your-api-key"
      }
    }
  }
}

OpenClaw

Add to your TOOLS.md:

### AgentTodo
- API: https://agenttodo.vercel.app/api
- Key: YOUR_API_KEY
- Auth: `Authorization: Bearer <key>`
- Create: `POST /api/tasks` (title required)
- Claim next: `POST /api/tasks/next`
- Complete: `POST /api/tasks/:id/complete`

OpenClaw agents will automatically use the API via curl.

---

Use Cases

Solo dev + AI agent

You plan the work, your agent executes. Create tasks from the dashboard, let your agent claim and complete them autonomously. Review results when you're ready.

Team + multiple agents

Multiple developers and agents share one queue. Each agent claims work based on its intents (build, research, test). No double-work — the queue handles coordination.

CI/CD integration

Trigger task creation from your pipeline. Failed deploy? Auto-create a blocked task. Tests pass? Mark the review task as done. Use curl or the API directly from GitHub Actions, GitLab CI, or any workflow.

# GitHub Actions example
- name: Create review task
  run: |
    curl -X POST https://agenttodo.vercel.app/api/tasks \
      -H "Authorization: Bearer ${{ secrets.AGENTTODO_KEY }}" \
      -H "Content-Type: application/json" \
      -d '{"title": "Review deployment ${{ github.sha }}", "intent": "review", "priority": 4}'

---

Self-Hosting

AgentTodo is a standard Next.js 14 app backed by Supabase.

Prerequisites

• Node.js 18+
• Supabase project (free tier works)
• pnpm (recommended) or npm

Setup

git clone https://github.com/EricStrohmaier/agenttodo.git
cd agentboard
pnpm install

Create .env.local:

NEXT_PUBLIC_SUPABASE_URL=https://your-project.supabase.co
NEXT_PUBLIC_SUPABASE_ANON_KEY=your-anon-key
SUPABASE_SERVICE_ROLE_KEY=your-service-role-key

Run database migrations:

npx supabase db push

Start the dev server:

pnpm dev

Deploy to Vercel

Add your Supabase environment variables in the Vercel dashboard and you're live.

---

Tech Stack

• Next.js 14 — App Router, API routes, server components
• Supabase — Postgres, Auth, Realtime subscriptions
• Tailwind CSS — Utility-first styling
• shadcn/ui — Component library
• TypeScript — End to end

---

Contributing

Contributions welcome! Please:

1. Fork the repo
2. Create a feature branch (git checkout -b feat/my-feature)
3. Commit your changes
4. Open a pull request

See CONTRIBUTING.md for detailed guidelines.

---

License

MIT

---

Built for agents, by humans (and agents). 🤖

agenttodo.vercel.app · GitHub