隐私政策

最后更新日期：2026年5月8日

上传的图片

When you use the extraction demo, your browser resizes the selected image and sends it to the ZapOCR Cloudflare Worker. The Worker stores the resized image in Cloudflare R2 for admin review, logs the requested field names, document type hint, custom instructions, model response, and extraction metadata in Cloudflare D1, and forwards the image to Gemini so the model can extract text.

存储

ZapOCR keeps extraction logs that include the resized uploaded image, extracted values, confidence scores, evidence notes, model name, timing, quota state, country code, user agent, and a pseudonymous requester identifier. Field definitions, such as Content or Invoice Date, are stored in your browser local storage so they can be reused on future visits.

免费扫描额度

为了执行 10 次免费扫描限制，ZapOCR 会使用请求信息的哈希值，例如 IP 地址和浏览器 user agent，在 Cloudflare Durable Objects 中保存匿名额度计数器。ZapOCR 不会保存原始 IP 地址。 如果你登录，Worker 会验证你的 Firebase ID token，并使用 Firebase 用户 ID 的哈希值保存登录后扫描使用量。

第三方处理

Gemini 处理上传的图片和提示词以返回 OCR 结果。除非您被允许将敏感数据发送给模型提供商，否则请勿上传包含此类数据的图片。

Payments and scan credits

If you buy scan packs, Stripe processes the payment and sends confirmation metadata to ZapOCR so paid scan credits can be added to your quota ledger.

联系我们

如有疑问，请发送至 [email protected].