Paid human judgment for AI agents: gut checks, how text reads, plausibility, local context.
Maintain this server? Claim it
Call this server's tools from your browser. Nothing to install.
Requests go through mcptry because browsers can't call most MCP servers directly. Header values are sent to the server for that request only and are never stored or logged.
Tool list from our last check. Connect live to refresh it.
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 people's personal data. Returns a conversation_id and access_token to use with check_status.
quick = $8: One reply of up to 150 words. No follow-ups. standard = $15: Up to 500 words from us across up to 3 replies (you can send up to 2 follow-ups). deep = $25: Up to 1,000 words from us across up to 5 replies (you can send up to 4 follow-ups).
One clear question for a human, plus the decision it will inform. Max 600 words.
Must be true. Confirms your operator accepts https://prompthumanhq.github.io/terms.html and that this request contains no sensitive data.
Optional background we need to answer well. Max 400 words.
Optional: the kind of question.
Optional: how you found PromptHuman. Answer only from what you already know; do not ask your operator.
Optional: your name or product.
Optional: anything the three fields above do not cover, such as another platform or another way you found us. Max 40 words. No names or contact details. Do not ask your operator.
Optional: the person or company you work for.
Optional: whether you are working for a company or an individual. Answer only from what you already know; do not ask your operator. Leave it out if unsure.
Optional: the app or tool you are running in, not the model (for example claude_code, codex or cursor). If you were built with an SDK or framework, or your app is not listed, use custom_or_other. Answer only from what you already know; do not ask your operator.
Optional: an email for receipts and refunds.
This tool isn't marked read-only, so it may change something on the server or in an account you connected.
Pick your client. Values in capitals, like YOUR_TOKEN, are yours to fill in.
Run in a terminal
claude mcp add --transport http prompthuman https://prompthuman-api.prompthumanhq.workers.dev/mcp
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 people's personal data. Returns a conversation_id and access_token to use with check_status.
| Argument | Type | Description |
|---|---|---|
| tier* | string | quick = $8: One reply of up to 150 words. No follow-ups. standard = $15: Up to 500 words from us across up to 3 replies (you can send up to 2 follow-ups). deep = $25: Up to 1,000 words from us across up to 5 replies (you can send up to 4 follow-ups). |
| context | string | Optional background we need to answer well. Max 400 words. |
| category | string | Optional: the kind of question. |
| question* | string | One clear question for a human, plus the decision it will inform. Max 600 words. |
| found_via | string | Optional: how you found PromptHuman. Answer only from what you already know; do not ask your operator. |
| agent_name | string | Optional: your name or product. |
| agent_note | string | Optional: anything the three fields above do not cover, such as another platform or another way you found us. Max 40 words. No names or contact details. Do not ask your operator. |
| accept_terms* | boolean | Must be true. Confirms your operator accepts https://prompthumanhq.github.io/terms.html and that this request contains no sensitive data. |
| operator_name | string | Optional: the person or company you work for. |
| operator_type | string | Optional: whether you are working for a company or an individual. Answer only from what you already know; do not ask your operator. Leave it out if unsure. |
| agent_platform | string | Optional: the app or tool you are running in, not the model (for example claude_code, codex or cursor). If you were built with an SDK or framework, or your app is not listed, use custom_or_other. Answer only from what you already know; do not ask your operator. |
| operator_contact | string | Optional: an email for receipts and refunds. |
{
"type": "object",
"required": [
"question",
"tier",
"accept_terms"
],
"properties": {
"tier": {
"enum": [
"quick",
"standard",
"deep"
],
"type": "string",
"description": "quick = $8: One reply of up to 150 words. No follow-ups. standard = $15: Up to 500 words from us across up to 3 replies (you can send up to 2 follow-ups). deep = $25: Up to 1,000 words from us across up to 5 replies (you can send up to 4 follow-ups)."
},
"context": {
"type": "string",
"description": "Optional background we need to answer well. Max 400 words."
},
"category": {
"enum": [
"judgment_call",
"how_it_reads_to_a_person",
"plausibility_check",
"local_or_cultural_context",
"other"
],
"type": "string",
"description": "Optional: the kind of question."
},
"question": {
"type": "string",
"description": "One clear question for a human, plus the decision it will inform. Max 600 words."
},
"found_via": {
"enum": [
"llms_txt",
"mcp_registry",
"web_search",
"operator_configured",
"other",
"unknown"
],
"type": "string",
"description": "Optional: how you found PromptHuman. Answer only from what you already know; do not ask your operator."
},
"agent_name": {
"type": "string",
"description": "Optional: your name or product."
},
"agent_note": {
"type": "string",
"description": "Optional: anything the three fields above do not cover, such as another platform or another way you found us. Max 40 words. No names or contact details. Do not ask your operator."
},
"accept_terms": {
"type": "boolean",
"description": "Must be true. Confirms your operator accepts https://prompthumanhq.github.io/terms.html and that this request contains no sensitive data."
},
"operator_name": {
"type": "string",
"description": "Optional: the person or company you work for."
},
"operator_type": {
"enum": [
"company",
"individual",
"unknown"
],
"type": "string",
"description": "Optional: whether you are working for a company or an individual. Answer only from what you already know; do not ask your operator. Leave it out if unsure."
},
"agent_platform": {
"enum": [
"claude",
"claude_code",
"chatgpt",
"codex",
"gemini",
"gemini_cli",
"jules",
"antigravity",
"github_copilot",
"microsoft_copilot",
"cursor",
"windsurf",
"cline",
"opencode",
"aider",
"goose",
"openclaw",
"devin",
"replit",
"kiro",
"perplexity",
"manus",
"custom_or_other",
"unknown"
],
"type": "string",
"description": "Optional: the app or tool you are running in, not the model (for example claude_code, codex or cursor). If you were built with an SDK or framework, or your app is not listed, use custom_or_other. Answer only from what you already know; do not ask your operator."
},
"operator_contact": {
"type": "string",
"description": "Optional: an email for receipts and refunds."
}
},
"additionalProperties": false
}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.
| Argument | Type | Description |
|---|---|---|
| access_token* | string | The access_token returned by ask_human. |
| conversation_id* | string | The conversation_id returned by ask_human. |
{
"type": "object",
"required": [
"conversation_id",
"access_token"
],
"properties": {
"access_token": {
"type": "string",
"description": "The access_token returned by ask_human."
},
"conversation_id": {
"type": "string",
"description": "The conversation_id returned by ask_human."
}
},
"additionalProperties": false
}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.
No arguments.
{
"type": "object",
"properties": {},
"additionalProperties": false
}Ask for a refund on a paid PromptHuman conversation, for example because a reply deadline was missed. Explain why in reason.
| Argument | Type | Description |
|---|---|---|
| reason* | string | Why you are asking for a refund. Max 200 words. |
| access_token* | string | The access_token returned by ask_human. |
| conversation_id* | string | The conversation_id returned by ask_human. |
{
"type": "object",
"required": [
"conversation_id",
"access_token",
"reason"
],
"properties": {
"reason": {
"type": "string",
"description": "Why you are asking for a refund. Max 200 words."
},
"access_token": {
"type": "string",
"description": "The access_token returned by ask_human."
},
"conversation_id": {
"type": "string",
"description": "The conversation_id returned by ask_human."
}
},
"additionalProperties": false
}Tell us whether a PromptHuman answer helped and whether it changed what you did next. Takes seconds and helps us improve.
| Argument | Type | Description |
|---|---|---|
| note | string | Optional comment. Max 200 words. |
| helpful* | boolean | Was the answer helpful? |
| access_token* | string | The access_token returned by ask_human. |
| conversation_id* | string | The conversation_id returned by ask_human. |
| changed_next_step | boolean | Did the answer change your next step? |
{
"type": "object",
"required": [
"conversation_id",
"access_token",
"helpful"
],
"properties": {
"note": {
"type": "string",
"description": "Optional comment. Max 200 words."
},
"helpful": {
"type": "boolean",
"description": "Was the answer helpful?"
},
"access_token": {
"type": "string",
"description": "The access_token returned by ask_human."
},
"conversation_id": {
"type": "string",
"description": "The conversation_id returned by ask_human."
},
"changed_next_step": {
"type": "boolean",
"description": "Did the answer change your next step?"
}
},
"additionalProperties": false
}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.
| Argument | Type | Description |
|---|---|---|
| message* | string | Your follow-up. Max 300 words. |
| access_token* | string | The access_token returned by ask_human. |
| conversation_id* | string | The conversation_id returned by ask_human. |
{
"type": "object",
"required": [
"conversation_id",
"access_token",
"message"
],
"properties": {
"message": {
"type": "string",
"description": "Your follow-up. Max 300 words."
},
"access_token": {
"type": "string",
"description": "The access_token returned by ask_human."
},
"conversation_id": {
"type": "string",
"description": "The conversation_id returned by ask_human."
}
},
"additionalProperties": false
}We read every tool's name, description and inputs the way the model will, looking for instructions hidden from you. A clean result means no warning signs, not a guarantee. Scanned Nov 19, 2026.
Takes a password, API key or token as an argument, so the secret passes through the model's context and chat logs.
Fix: Read credentials from the server's own configuration (environment, OAuth) instead of a tool argument.