Privacy Policy

This policy explains what Code By AI stores, what it does not store, and who else processes your data.

Last updated: November 12, 2026

Your Prompts, Code and Model Output

We do not store the content of your requests. Prompts, code, files, images and model responses pass through our gateway and are not saved to our database or used to train or improve anything.

  • To produce a response, request content is forwarded to the upstream model service provider for the model you chose. That provider processes it under its own terms. If no upstream channel is available for the model you asked for, we refuse the request instead of answering it with a different model.
  • Generated images are returned to you directly and are not kept on our servers.
  • Error logs record the status code, the upstream's machine-readable error type and the size of the failing response — enough to diagnose a failure without keeping your content. They do not record request content.

What We Store

We keep only what is needed to run your account and bill usage:

  • Email abuse protection: Cloudflare Turnstile verifies requests to send registration, verification, password-reset and email-login emails. Cloudflare processes browser and network signals. We store a hash of the verification token to prevent reuse; these hashes expire after a few minutes and are removed during routine rate-limit cleanup. We do not store the raw Turnstile token.
  • Onboarding reward protection: Cloudflare Turnstile verifies claims from the API-key dialog. We keep the reward amount, time, linked account/key, a verification-token hash and keyed hashes of the email and network address to prevent repeated or automated claims. Reward audit hashes remain after account deletion to prevent re-registration from claiming again. Raw verification tokens and network addresses are not stored in this reward audit.
  • Account: email address, optional name and avatar, a password hash if you set a password, and which Google or GitHub account you signed in with.
  • Balance and payments: your balance, top-up amounts, currency and payment status. Card and wallet details are handled by Stripe; we never receive them.
  • Usage: token counts and cost per API key, per model and per day.
  • API keys: the keys you create and their names.

We use this data to provide the service, bill usage, prevent abuse and contact you about your account. We do not sell it.

IP Addresses and Site Analytics

  • Your IP address is used for rate limiting and regional access checks. Rate-limit records expire with their time window.
  • Our own site analytics store a salted daily hash of the IP address, never the address itself, together with the page visited, referring site, campaign tags, country and device type.
  • To measure our sign-up funnel, we store a stable one-way hash of your account ID with conversion events: creating an API key, your first API request and each top-up. The same hash is used every time, so these events can be linked to one another, but it cannot be turned back into your ID without our private salt.
  • We do not use third-party advertising or analytics trackers.

Cookies and Browser Storage

  • Sign-in cookies: keep you logged in and protect sign-in forms.
  • Registration attribution: a signed first-party cookie lasts up to 7 days and records the first recorded visit’s channel, landing page and time. When you create an account, we save this snapshot with the account to measure registration and activation by source. Later sign-ins and payments do not change it. It contains no API keys, prompts or raw IP addresses.
  • Preference cookie: remembers whether the dashboard sidebar is open.
  • Browser storage: a random per-tab identifier for site analytics, and short-lived state such as a pending payment. The per-tab identifier is included with the payment we send to our payment processor, so that a top-up can be attributed to the visit that produced it; it is not used to identify you and is not shared with anyone else.

Service Providers

  • Upstream model service providers: receive request content to generate responses, as described above.
  • Stripe: processes card, Alipay and WeChat Pay payments.
  • Feishu (Lark): receives internal operational alerts that only our operations team can see. Payment alerts contain your user ID, the amount and the currency. Refund alerts contain your user ID, the balance reversed and the Stripe payment ID. Dispute alerts contain the Stripe dispute ID, status, amount and currency. Invitation-reward alerts contain the inviter's and the invitee's user IDs. Billing-failure alerts contain your user ID, the model and the amount. Alerts never contain your email address, payment method or request content.
  • Google and GitHub: sign-in, only if you choose them.
  • Resend: sends sign-in and account emails.
  • Cloudflare: hosts the site and gateway.

These providers may process data in countries other than your own.

Retention and Your Rights

  • Account, balance and usage data are kept while your account exists. Payment records may be kept longer where accounting or legal rules require it.
  • Visit and conversion event records are deleted after 400 days. The registration source snapshot is part of your account record and is removed when the account is deleted.
  • You can ask for a copy of your data, correct it, or have your account and its data deleted.

To make a request, email cfjwlchangji@gmail.com from the address on your account. We usually reply within 24 to 48 hours.

Changes to This Policy

We will post any change on this page and update the date above. For material changes we will also notify you by email.

Questions about this policy: cfjwlchangji@gmail.com