MCP Human Assistant

Example Use Cases

• 1

  Get short text answers from real humans for tasks that require common sense or subjective judgement by creating MTurk HITs.

• 2

  Validate or augment generated content by sending a question to workers and retrieving their submitted responses via a hosted form.

• 3

  Monitor HIT lifecycle and account info by checking HIT status, listing HITs, and retrieving MTurk account balance.

Description

This repository implements a Model Context Protocol (MCP) server that integrates with Amazon Mechanical Turk to create ExternalQuestion HITs and collect worker responses via a hosted form. It lets AI agents ask humans questions, poll for assignment results, and check HIT/account status (including balance and list of HITs). It is intended as a proof-of-concept for human-in-the-loop workflows and supports sandbox testing by default.

Capabilities(7 total)