Privacy

Privacy

Short version: we keep as little as we can, we never sell anything about you, and the prompt you type is deleted after 90 days.

What happens to the prompt you type

When you press Enhance, your draft is sent to our server so we can work out which template fits. It is converted into a numeric representation by Google's Vertex AI, which processes the text to do so. That is the only outside service involved in the normal path.

If you press try harder, your draft is additionally sent to an AI provider — Fireworks or Anthropic — to make a better guess. That is the only time your text goes to one of those companies, and it only happens because you asked.

Your draft and the prompt we generate for you are stored for 90 days, then automatically deleted. We keep the rest of the record — which template you used, which options you picked, which model you chose, and whether you copied the result — for longer, because that is what tells us which templates work. That part contains none of your writing.

We do not train models on your prompts. We do not sell them. Nobody outside the company reads them except as described above.

What we deliberately do not have

• No email address, name or photo in our database. If you sign in, your email lives with our sign-in provider (Firebase Authentication) and is never copied into our own records.
• No password, anywhere. You sign in with Google, Apple, or a link we email you. We never receive a password, so we cannot lose one.
• No IP addresses in our analytics, and no full browser fingerprint — we record only the browser family, like "Chrome".
• No cross-site tracking, and no connection between your device identifier and any advertising identifier. Ever.

The device identifier

Your browser and app are given a random identifier so we can apply rate limits and count sessions without knowing who you are. It is not linked to your name, your email, or any advertiser. If you later create an account, we record that the two are the same person so your history is yours — that link is the only thing that connects them.

Your account

An account exists for exactly one reason: to keep your saved templates across devices. It holds your sign-in identifier, your default model, and your language. That is all.

What is public

Presets you publish are public — their title, description, tags, and any text you saved in their fields, along with your handle. Your drafts are never published. See the content policy.

Stored on your device, not ours

Your recent templates, your appearance and language preferences, and the draft being handed between screens stay in your browser's local storage. They never reach us.

Your controls

• Clear recents — on Mine, at any time, no account needed.
• Delete account and data — on Me. It removes your account, your saved templates, and the text of every prompt you have typed, including from before you signed in. It tells you exactly what will go before you confirm.
• Your presets stay published when you delete your account — the byline becomes deleted account. Other people have saved and built on them. If you want one gone, unpublish it first; it then leaves the library and search immediately, though anyone who already saved it keeps their copy.

Who else receives data

Who	What they get	Why
Google Cloud (Vertex AI)	your draft, to match a template	every Enhance
Fireworks or Anthropic	your draft	only when you press "try harder"
Firebase Authentication	your email or sign-in identity	only if you sign in
Google	hosting and infrastructure logs	running the service

We do not currently show advertising. When we do, it will appear only on library pages and never inside the flow, and it will never be connected to your device identifier.

Children

PromptPiglet is not directed at children under 13 and we do not knowingly collect their data.

Changes, and how to reach us

If we change what we collect, we will say so on this page and date it. Questions, or a request to see or delete your data: privacy@promptpiglet.com.

PromptPiglet is operated by David Li, an individual, from Hong Kong SAR, who is responsible for the data described here under the laws of Hong Kong SAR.