# PromptHuman

> PromptHuman is a paid MCP server for human-in-the-loop tasks. Pay-per-question written answers from real people, for AI agents needing human judgment. It runs as a hosted remote server, so there is nothing to install locally.

- **Kind:** MCP server
- **Pricing:** Paid
- **Platforms:** API
- **Website:** https://prompthumanhq.github.io
- **MCP Registry name:** io.github.prompthumanhq/prompthuman
- **Categories:** [human-in-the-loop tasks](https://ai-powered.com/categories/human-in-the-loop-tasks)
- **Listing updated:** 2026-11-20

## About

PromptHuman lets an AI agent send a question to real people and receive a written reply. The kinds of questions it covers are gut checks, how a message or design will read to people, plausibility checks, local norms and etiquette, and tie-breaking between options.

The agent connects through an MCP server or a JSON API. Requests are accepted or declined within 24 hours, and payment via a Stripe link is only requested after acceptance. The site lists three tiers: a single reply of up to 150 words for $8, up to 500 words across three replies for $15, and up to 1,000 words across five replies for $25.

The service is aimed at developers building agents that need a human perspective on subjective matters. It explicitly excludes legal, medical, tax and financial advice, credential handling, taking actions on a user's behalf, and bypassing CAPTCHAs or verification checks.

## Features

- **Written human replies:** Responses come from real people in writing, with a stated first-reply window of 24 hours after payment.
- **Pay per question:** Each request is prepaid by card through a Stripe link and is only charged after the request is accepted.
- **Tiered reply lengths:** Quick take, Standard and Deep dive tiers set word limits and the number of follow-up replies.
- **Refund on missed deadline:** The site states a full refund applies if a reply deadline is missed.
- **MCP and JSON access:** Agents can connect through a streamable HTTP MCP server or a JSON API.
- **Data retention limits:** Conversations are deleted 90 days after they close and are stated not to be used for model training.

## Good for

- Checking whether an email or message will read well to a real person before sending
- Getting a gut check on a design, name or gift idea
- Sanity-testing whether a plan or behaviour seems plausible to ordinary people
- Learning local etiquette and norms before a trip or meeting
- Breaking a tie between two options a person would be likely to choose

## Install

### Claude Code (Terminal)

```sh
claude mcp add --transport http prompthuman https://prompthuman-api.prompthumanhq.workers.dev/mcp
```

### Claude Desktop (Settings → Connectors → Add custom connector)

```sh
https://prompthuman-api.prompthumanhq.workers.dev/mcp
```

### Cursor (~/.cursor/mcp.json)

```json
{
  "mcpServers": {
    "prompthuman": {
      "url": "https://prompthuman-api.prompthumanhq.workers.dev/mcp"
    }
  }
}
```

### VS Code (.vscode/mcp.json)

```json
{
  "servers": {
    "prompthuman": {
      "type": "http",
      "url": "https://prompthuman-api.prompthumanhq.workers.dev/mcp"
    }
  }
}
```

### Windsurf (~/.codeium/windsurf/mcp_config.json)

```json
{
  "mcpServers": {
    "prompthuman": {
      "serverUrl": "https://prompthuman-api.prompthumanhq.workers.dev/mcp"
    }
  }
}
```

### Grok Bot (Grok Bot desktop app → Connect apps → Add a custom connector)

```sh
https://prompthuman-api.prompthumanhq.workers.dev/mcp
```

Enter a Name (for example "PromptHuman"), choose Type: Remote HTTPS, paste this as the Server URL and choose Add plugin.

### Other clients (your client's MCP settings (most use this shape))

```json
{
  "mcpServers": {
    "prompthuman": {
      "type": "http",
      "url": "https://prompthuman-api.prompthumanhq.workers.dev/mcp"
    }
  }
}
```

## Tools

2 of 6 tools say they're read-only. 4 of 6 tools say they can delete or change data. From the server's own tool annotations; not verified by us.

- `get_prompthuman_info` (says read-only): Get PromptHuman's tiers, prices, word limits, turnaround times, refund rules, out-of-scope list and sensitive-data rules. Call this before your first ask_human.
- `ask_human` (says it can change data): Ask a real person for human judgment when you need it: a gut check, how something will read to a person, whether a plan is plausible, local or cultural context, or a tie-breaking call. Paid and asynchronous: we review within 24 hours, you pay only after we accept, and the first reply arrives within 24 hours of payment. Never include passwords, keys, card or ID numbers, health records, or other peo
- `check_status` (says read-only): Get the status of a PromptHuman request: whether it was accepted, the payment link, reply deadlines, the full message thread, and how many words and replies remain. Poll at most once every 10 minutes.
- `send_followup` (says it can change data): Send a follow-up message in a PromptHuman conversation after a human has replied. Only works when check_status shows can_send_followup: true. Max 300 words; no sensitive data.
- `request_refund` (says it can change data): Ask for a refund on a paid PromptHuman conversation, for example because a reply deadline was missed. Explain why in reason.
- `send_feedback` (says it can change data): Tell us whether a PromptHuman answer helped and whether it changed what you did next. Takes seconds and helps us improve.

## FAQ

### How much does a question cost?

The listed prices are $8 for one reply of up to 150 words, $15 for up to 500 words across three replies, and $25 for up to 1,000 words across five replies.

### What kinds of questions are excluded?

The site excludes legal, medical, tax, financial, investment and engineering advice, access to accounts or credentials, actions on a user's behalf, solving CAPTCHAs, and deception, impersonation, harassment or illegal requests.

### Is payment taken before a reply?

No. Nothing is charged until the request is accepted, and the agent then pays a Stripe link within 48 hours.

### How is sensitive data handled?

Passwords, API keys, card or bank numbers, government IDs, health records and personal information about third parties are not accepted, and the server rejects obvious cases automatically.

### How do I add PromptHuman to Claude Code?

Run this command in your terminal: claude mcp add --transport http prompthuman https://prompthuman-api.prompthumanhq.workers.dev/mcp

### What are the best alternatives to PromptHuman?

The closest alternatives to PromptHuman are hands.txt: a human proxy for AI agents and Human Use.

## Where this comes from

- Description: written from the project's website and its MCP Registry entry.
- Status and tools: checked automatically by AIPoweredBot on 21 November 2026.
- Install steps: from the official MCP Registry entry.

## Alternatives

- [hands.txt: a human proxy for AI agents](https://ai-powered.com/mcp/hands-txt-a-human-proxy-for-ai-agents): hands.txt: a human proxy for AI agents is a paid MCP server for human-in-the-loop tasks.
- [Human Use](https://ai-powered.com/mcp/human-use): Human Use is an open-source MCP server for human-in-the-loop tasks.

---
Source: https://ai-powered.com/mcp/prompthuman
