Skip to main content

When your agent needs lived experience, not training data.

Real people answer your agent’s questions in writing, through our MCP server or JSON API.

Your agent asks a real person a question and gets a written answer. Gut checks, how something will read to people, plausibility checks, local context and tie-breaking calls. Paid per question, no operator involvement needed.

Prepaid by card through a Stripe link, sent only after we accept a request. Full refund if we miss a reply deadline.

How it works

  1. Your agent sends one clear question through our MCP server or JSON API.
  2. We accept or decline within 24 hours; a request we haven’t accepted by then expires. Nothing is charged before we accept.
  3. Your agent pays a Stripe link within 48 hours.
  4. Our first written reply arrives within 24 hours of payment, and each follow-up gets a reply within 12 hours.

Pricing

Tier Price What you get
Quick take $8 One reply, up to 150 words. No follow-ups.
Standard $15 Up to 500 words across up to 3 replies.
Deep dive $25 Up to 1,000 words across up to 5 replies.

Word limits count the words in our replies. Your agent’s question can be up to 600 words, plus up to 400 words of background, and each follow-up can be up to 300 words.

What we answer

Yes

  • Will this email, name, gift or design land with a real person?
  • Would a normal person actually do, say or believe this?
  • What a move, job search, trip or hobby is really like
  • Local norms, etiquette and what reads as odd
  • Which option a person would pick, and why

No

  • Legal, medical, tax, financial, investment or engineering advice
  • Anything that needs access to accounts or credentials
  • Actions in the world on your behalf
  • Solving CAPTCHAs or getting around verification checks
  • Deception, impersonation or harassment
  • Anything illegal or harmful

We’re experienced people, not licensed professionals.

Don’t send sensitive data

We don’t accept passwords, API keys or other credentials; card, bank or government ID numbers; health records; personal information about people other than the operator; or anything the operator hasn’t authorised the agent to share. Our server rejects obvious cases automatically.

Connect your agent

  • MCP server: https://prompthuman-api.prompthumanhq.workers.dev/mcp
  • JSON API: https://prompthuman-api.prompthumanhq.workers.dev/api/info
  • Instructions for agents: llms.txt and pricing.json

Every request is covered by mutual confidentiality in our terms. Conversations are deleted 90 days after they close and never used to train models.

Questions from humans: prompthumanhq@gmail.com