PromptPrio

The task manager for ChatGPT

Give ChatGPT a real task queue. TaskPrio exposes one global priority list over a clean REST API (and MCP), so ChatGPT — via a custom GPT or a direct API call — pulls the next task, works it, and reports back. You set the order; it runs the list.

Open the board — free →

Option A — a custom GPT (no code)

  1. Create a GPT in ChatGPT → Configure → Actions.
  2. Import the schema from https://promptprio.com/api/openapi.json.
  3. Set authentication to Bearer and paste your TaskPrio API token (from the in-app account menu).
  4. Tell the GPT: "Work my TaskPrio queue: call get_next_task, do it, complete_task, repeat."

ChatGPT now reads and completes tasks from the same queue you see on the web board.

Option B — the REST API directly

GET  /api/next                       # → highest-priority task (Bearer token)
# …do the work…
POST /api/tasks/{id}/complete        # { "result": "what got done" }
# repeat until { "task": null }

This is the same loop your agent runs over MCP — just over HTTP. Full details in the API reference and the machine-readable OpenAPI spec.

Why route ChatGPT through a queue

Frequently asked questions

How do I connect ChatGPT to a task queue?

Either import TaskPrio's OpenAPI spec as a custom-GPT Action with a Bearer token, or call GET /api/next + POST /api/tasks/{id}/complete directly. MCP-capable clients can use the bundled server too.

Can ChatGPT work my tasks automatically?

Yes — with the Action wired up it loops get_next_task → work → complete_task over your one priority queue.

Do I need a paid plan?

Custom GPTs/Actions need a paid ChatGPT plan from OpenAI; TaskPrio is free (Pro $10/mo). The REST path works from any HTTP-capable setup.

Start free — no card →   How the queue works →