Skip to main content

Privacy

Effective November 19, 2026. Last updated November 19, 2026.

In short

  • We collect what we need to answer requests, deliver written replies and take payment.
  • Conversations (requests, follow-ups and our replies) are kept separately, read by us and our consultants, and deleted 90 days after the conversation closes.
  • For the long term, we keep only a de-identified service log: a short summary, in our own words, of each question we accept, with no names, contact details or company names, plus administrative details such as dates and tier, and a client code that links one payer’s requests without recording who they are.
  • We never log the content of out-of-scope requests.
  • We don’t sell your information, show ads, or use your conversations to train AI models.
  • Questions or deletion requests: prompthumanhq@gmail.com. Send a deletion request from the email address you paid with.

Who this covers

PromptHuman (“we”, “us”) connects AI agents with real people for short, paid questions answered in writing. This policy explains how we handle personal information across everything we do: our website, our MCP server and JSON API, the requests we receive, the replies we write and the payments we take.

“You” means the operator of the requesting agent (the person or company that runs it), including when the agent contacts us on your behalf. We treat requests from an agent as coming from its operator. This policy works alongside our Terms, including their confidentiality section. Where the two differ, the one that gives your information more protection applies.

What we collect

Requests and conversations. Agents submit requests only through our MCP server or JSON API. We receive the question, any background, the tier chosen, follow-up messages, and technical details needed to run the service, such as when each message arrived. An agent may also tell us its own name, your name and a contact email address; all three are optional. We write replies that we return to your agent through the same channel.

Optional details about the agent. An agent may also tell us, by choosing from fixed lists, whether it works for a company or an individual, what platform it runs on, and how it found us. It may add a short note in its own words. All of these are optional, and we ask agents to answer only from what they already know. The three fixed-list answers go into the service log described below. The note stays with the conversation and is deleted with it.

Internet addresses. To limit abuse, our server keeps a scrambled fingerprint of the internet (IP) address each new request comes from, and uses it only to cap how many requests one address can send per hour. The fingerprint can’t be turned back into the address, we never store the address itself, and the fingerprint is erased about a day after the request. Cloudflare, as our host, also keeps short-lived technical logs of requests to our server (not their content) under Cloudflare’s privacy policy.

Email. If a person emails us with a question about the service or a refund, we receive the sending email address and the message.

Payments. Requests are paid by card through Stripe. Through Stripe we receive your email address, your name or business name if you provide one, your billing country and postal code, the request ID attached to your payment link, whether you ticked the box agreeing to our Terms, and the amount and status of the payment. We never see or store full card numbers; card details are handled by Stripe and never reach us. Our server, hosted by Cloudflare, keeps Stripe’s reference for your payment and the dates it was paid or refunded. When a payment is confirmed, our server uses the email address once to work out a client code (see The service log) and does not store the address itself.

Phone and voicemail. If you call or text our business number, we receive your number and any message you leave.

Our website. Our website is a static site hosted by GitHub Pages. We don’t use cookies, analytics or tracking scripts. GitHub, as our host, may log technical information such as visitors’ IP addresses for security purposes, under GitHub’s privacy statement.

What we ask you not to send. As our Terms say, never send passwords, API keys, tokens or other credentials; payment card, bank account or government ID numbers; health or medical records; personal information about anyone other than the operator; or information you are not authorised to share. Our server rejects obvious cases automatically, and we may decline or close a request that contains such information and delete it without answering. We also decline requests for legal, medical, tax, financial, investment or engineering advice as out of scope.

How we use it

  • To decide whether a request is in scope and reply to it.
  • To write and deliver replies and follow-up replies.
  • To take payments, issue refunds and keep accounting and tax records.
  • To keep the de-identified service log described below, which we use to improve our service.
  • To prevent fraud and abuse, enforce our Terms and meet legal obligations.

We don’t sell personal information, use it for advertising, or use conversations to train AI models.

How we store and delete it

We keep three kinds of records, each stored and kept for a different period.

Record What it contains How long we keep it
Conversations Requests, background, follow-ups and our replies Deleted 90 days after the conversation closes
Payment records Payment and refund details held by Stripe (under its own policies) and in our accounts As long as tax and accounting law requires (generally up to seven years)
Service log De-identified summaries and administrative details Kept indefinitely, unless you ask us to remove your entry

Conversations travel over encrypted connections and are stored with our hosting provider, Cloudflare, separately from the service log. They are read by us and our consultants; Cloudflare can technically access what it stores for us. We share them with no one else, except as described under Who we share it with. Our phone notifications about new requests contain a request ID and tier, never request content. We keep a conversation past 90 days only if we need it to resolve a dispute, refund or chargeback, or if the law requires us to keep it, and we delete it once that need ends.

When a conversation is deleted, automatically or at your request, it is removed from our live database straight away. Our database provider keeps rolling backups for a short period (currently up to 7 days), so a deleted conversation can remain in a backup for that long before it is gone for good.

Your agent’s platform may keep its own copy of the conversation. That copy is your responsibility, and deleting ours doesn’t affect it.

The service log

To understand what agents need and improve our service, we keep a log of every request. It contains:

  • for each request we accept, a short summary of the question (up to 40 words), written by one of us in our own words, leaving out names, contact details, company names and identifying specifics. Our server refuses a summary that contains an email address, a web link, a phone number or a phrase copied from the request;
  • a request ID and, for paid requests, a client code. The client code is made by scrambling the email address you paid with, using a secret key that only our server holds. We store the code, never the address. The same address always gives the same code, which lets us recognize repeat clients without recording who they are;
  • administrative details, such as how the agent found us, the agent platform, the type of operator (company, individual or unknown), the question category, the tier and price, the dates and times a request was received, reviewed and answered, the number of replies and words used, payment and refund status, and the agent’s feedback on whether our answer was helpful and whether it changed the agent’s next step.

The log never contains raw request text, names, email addresses, phone numbers, company names or payment details. If we decline a request, the log records only that it was declined and which general reason we chose, such as “out of scope” or “unclear”, never what the request said or any note we sent back.

A client code can’t be turned back into an email address. It can be matched to you only by someone who has both our secret key and your email address, which is how we find your records when you ask us to delete them. Because all of one payer’s entries carry the same code, those entries are linked to each other.

We and our consultants download copies of the log to study it. Our server keeps a record of every download (who, when and which file), so that if you ask us to remove your entries we can also remove them from those copies.

We may share aggregate statistics from the log, such as how many requests fall into each category, but never in a way that could identify a client or a specific request.

Who we share it with

  • Our consultants, who are bound to keep conversations confidential. Any of them can read the conversations on our private dashboard.
  • Service providers who store or process information for us: Cloudflare (our MCP server, JSON API and conversation storage), Stripe (payments), Telegram (phone alerts to our consultants, which carry a request ID, tier and a link to our private dashboard but never request content), Google (email and our business phone number) and GitHub (website hosting). Each handles information under its own privacy policy.
  • Legal and safety reasons, when we believe in good faith that the law requires it, or to protect the rights, property or safety of PromptHuman, our clients or others.
  • A business transfer, if PromptHuman is merged, sold or reorganized. Anyone who takes over must protect your information at least as well as this policy does.

We don’t share or publish conversations with anyone else.

Information about other people

A request may mention people other than you, such as your users or colleagues. You’re responsible for having the right to share their information with us. We treat it the same way as yours: it stays in the conversation record, it’s deleted 90 days after the conversation closes, and it never enters the service log.

Your choices and rights

You can ask us to:

  • tell you what personal information we hold about you and give you a copy;
  • correct information that’s wrong;
  • delete your conversations before the 90 days are up;
  • remove the entries for your requests from our service log;
  • stop using your information for a particular purpose.

Email prompthumanhq@gmail.com from the email address you paid with. That address is how we find your records and confirm they are yours. We’ll reply to it and ask you to confirm before we delete anything, because deletion can’t be undone. If you never paid, or can no longer use the address you paid with, include your conversation ID instead.

Tell us which you want: only your conversations and messages deleted, or your service log entries removed as well. If the address you paid with is shared, for example a company’s accounts mailbox, a request from it covers everyone who paid through it. We can’t delete a conversation while a refund or a payment on it is still being settled; we’ll tell you if that applies and delete it once it is.

We’ll respond within 30 days. We may need to keep payment records the law requires us to keep. The service log holds no names, contact details or request text. Its entries are tied to you only through the client code, so we find them from your paying address, or from a conversation ID, and remove them if you ask.

Security

We protect our accounts with two-step verification, encrypt conversations in transit, keep them separate from the service log, limit access to us and our consultants, and delete them on schedule. No system is perfectly secure. If a breach affects your personal information, we’ll notify you as the law requires.

International clients

We’re based in Pennsylvania, in the United States, and our service providers store information in the United States and elsewhere. If you contact us from another country, your information is transferred to and handled in the United States.

If you’re in the European Economic Area or the United Kingdom, we rely on these legal bases: performing our agreement with you (handling requests, replies and payments), our legitimate interest in improving our service (the de-identified service log) and in preventing fraud, and legal obligations (tax and accounting records). You also have the right to complain to your local data protection authority.

Children

Our service is for businesses and adults. We don’t knowingly collect information from anyone under 18.

Changes

If we change this policy, we’ll post the new version here and update the date at the top. If a change materially affects how we handle information we already hold, we’ll highlight it on this page before it takes effect.

Contact

prompthumanhq@gmail.com