Privacy Policy

To run your account and AI features, we collect and store:

• Account profile — your name, email, and avatar as provided by your OAuth sign-in (Google or GitHub)
• Your content — prompts, instruction sets, agent architectures, templates, codebase analyses, and knowledge-vault notes you create or save
• GitHub App installation — when you connect a repository, we store the installation id and the GitHub account login and account type it was installed on, to bind repository access to your claim of that installation
• Usage records — counts and token totals of your AI generations, used to enforce free and paid quotas
• Technical data — standard server logs (IP address, timestamps) and one sign-in cookie (see §4)

We do not run third-party analytics or advertising trackers, and we do not sell your data. Account and content records are stored in a managed Postgres database (Neon); saved codebase archives and their extracted files are stored in object storage (R2); and session and operational state is stored in KV.

Alongside your account, Prompterjack uses your browser's localStorage for on-device convenience:

• Drafts & anonymous work — so you don't lose work created before signing in
• Theme preference — your light/dark mode choice
• Architecture snapshots — temporarily cached for cross-page features like the OpenClaw setup guide

This on-device data stays in your browser. You can clear it at any time through your browser's settings (Clear Site Data). Provider API keys are handled separately and kept in your browser — see §3.

When you use an AI feature, your request goes to our Cloudflare Worker. Most generations run on Cloudflare Workers AI (the managed models built into our platform). Pro accounts may also be offered a frontier model that runs through Cloudflare's AI Gateway on OpenAI's infrastructure, billed to the platform — availability of that route is operator-controlled and is not something every request uses. Separately, the Worker's /api/generate endpoint still accepts a bring-your-own-key request for Anthropic, Google, or Groq through an AI Gateway and relays it to your chosen provider when present, but no bring-your-own-key model is currently selectable in the model picker, so the shipped app never routes a request that way today. The result is streamed back to you either way.

• Free and Pro generations run on Cloudflare Workers AI, or — for Pro accounts when the operator has enabled it — the OpenAI frontier route through Cloudflare's AI Gateway, using our platform account either way; no external provider key is involved. We record token usage to enforce quotas, but we do not store your prompt text for analytics or model training
• If a bring-your-own-key model is enabled, the key would stay in your browser (localStorage) and be sent with each request only to call your chosen AI provider through our Worker — never stored on our servers or logged; today no bring-your-own-key model is currently selectable, so this path is not reachable in the shipped app
• Each AI provider has its own privacy policy governing the data it receives

Prompterjack uses exactly one cookie: __Secure-prompterjack.session-token, a secure, HTTP-only cookie that keeps you signed in. It is strictly necessary to operate the service and is never used for advertising or cross-site tracking. We do not use:

• Google Analytics or any third-party analytics service
• Tracking pixels or web beacons
• Fingerprinting or device identification
• Social media tracking scripts
• Advertising networks

Because the sign-in cookie is essential, it works without a consent banner. Sign out, or clear site data, to remove it.

Prompterjack relies on the following third-party services, which act as our sub-processors:

• Cloudflare Pages — hosts the web app. Cloudflare may collect standard server logs (IP addresses, request timestamps). See Cloudflare's privacy policy.
• Cloudflare Workers & KV — run our API and proxy AI requests; KV holds session and operational data.
• Neon — managed Postgres database storing your account and saved content.
• Google & GitHub (OAuth) — authenticate your sign-in; we receive your basic profile (name, email, avatar).
• Cloudflare Workers AI — runs most generations on Cloudflare's managed models. The Worker also supports routing bring-your-own-key requests to Anthropic, Google, or Groq instead, each with its own data-handling policy, but no bring-your-own-key model is currently selectable, so the shipped app never routes a request there today.
• OpenAI (via Cloudflare AI Gateway) — receives prompt content for the Pro-only frontier model route when the operator has it enabled; billed to the platform through Cloudflare's Universal Endpoint, with no Prompterjack-held OpenAI key.
• Stripe — processes payments for paid plans. We never see or store your card details.
• Resend — sends transactional email (the optional weekly Pro fleet digest). Receives your email address and a summary of your monitored codebases' health; you can unsubscribe from any digest.

Several of these providers (including Cloudflare, Neon, and Stripe) operate on infrastructure located in the United States, so your personal data may be transferred to and processed there. We rely on the data-processing terms and transfer safeguards each provider makes available under their own agreements; you can review each provider's Data Processing Agreement on their website.

Prompterjack is an AI system. Some of its tools send what you provide to a language model and return that model's output to you — you are interacting with an AI system, not a person, and nothing is reviewed by a human before you see it. Output from those tools is artificially generated content and is labeled as AI-generated where it appears.

Not every tool uses a model: the security scanners and the codebase analyzer are rule-based, and their results are not AI-generated content. The Disclaimer lists every tool and says which is which.

We do not train, fine-tune, or host the underlying models, and we never train models on your prompts. The inference providers that run them are listed as sub-processors above.

If you are in the EU or UK, we process your personal data under the following legal bases in the GDPR:

• Performance of a contract (Art 6(1)(b)) — creating and running your account, signing you in, and producing the AI generations you request.
• Legitimate interests (Art 6(1)(f)) — keeping usage and quota tallies, maintaining security, preventing abuse, and operating the service reliably.
• Consent (Art 6(1)(a)) — optional communications such as the weekly Pro fleet digest, which you can withdraw at any time by unsubscribing.
• Legal obligation (Art 6(1)(c)) — where we are required to keep limited records, such as payment records held via Stripe for accounting purposes.

Data controller. Prompterjack is the data controller for the personal data described in this policy. You can contact us at chinmaypurohit1010@gmail.com.

Data retention. We keep your account and content data for as long as your account is active. Deleting an individual codebase yourself (Codebase → Delete) starts removing its stored objects immediately, on a best-effort basis — a failed batch is logged and left for the daily sweep to retry, and the request still confirms success either way. When you delete your whole account (§7), your account and database records are removed immediately and irreversibly; any remaining codebase archives and extracted files held in object storage are removed by a daily sweep — an object becomes eligible once it is at least 24 hours old, the sweep visits at most 1,000 candidate prefixes per run, and no fixed completion deadline is promised. Server logs are retained by our infrastructure providers according to their standard retention periods.

You are in control of the data we hold about you:

• Access / export — download the personal data we hold as a JSON file from Profile → Danger Zone → Export my data.
• Erasure — permanently delete your account and all associated data (codebases, analyses, teams you own, usage records, and every active session) from Profile → Danger Zone → Delete Account. Your account and database records are removed immediately and irreversibly; codebase archive and extracted-file objects held in object storage are removed by a daily sweep, once each object is at least 24 hours old, with no fixed completion deadline promised. Deleting an individual codebase yourself, rather than the whole account, starts removing its objects immediately on a best-effort basis, and the daily sweep clears anything left behind. An active subscription is cancelled as part of the deletion.
• Rectification — update your display name and profile from your Profile page.

If you can't reach your account, email chinmaypurohit1010@gmail.com and we will action the request.