Privacy Policy

1. Data we collect

The app indexes only the folder you select during onboarding through the system folder picker dialog (the app does not index other files by default). The folder is scanned recursively for image files with extensions: png, jpg, jpeg, heic, webp.

Images you index (content): when indexing, the app reads the image bytes from your selected folder and uploads them to our backend for processing (OCR, captioning, and embeddings).

Derived data: the app and backend generate and use OCR text (including an English translation for search), captions, and vector embeddings for search. On your device, this derived data is stored in a local SQLite database.

Chat data: when you use the app chat, we send your messages and relevant text (and, if vision is used in the chat, image attachments prepared by the app) to our backend for AI responses. Chat history is stored locally in a separate SQLite database on your device.

Local device identifier: the app generates a pseudonymous device_id (UUID), stores it in your macOS Keychain, and uses it to authenticate with the backend.

Usage and quota data: our backend tracks indexing and chat quotas per device (counts, daily reset timers, credits/spend). The app also sends your current timezone in an HTTP header to align daily limits.

Feedback data (if you send feedback): the backend accepts /v1/feedback requests containing a message, an optional reply email, optional app version and device fields, and up to 3 image attachments (base64) for debugging.

Subscription validation data: for paid features, the backend verifies StoreKit transactions via /v1/billing/verify.

Website: our marketing site at screenshotai.org is a static site. We use privacy-focused analytics (Cloudflare Web Analytics) for aggregated page views and performance metrics. It does not use advertising cookies and does not build ad profiles. If you email us or send Feedback from the app, we process that correspondence.

2. How we process your data

Indexing: the app uploads your image(s) to the backend. The backend resizes/compresses the image and then uses an AI provider accessed via OpenRouter to perform:

• OCR (extract text from image)
• Captioning (generate a descriptive caption)
• Visual embeddings (vector representation of the image)

After OCR is complete, OCR text may be translated to English for search indexing. Text embeddings are computed on our backend using an on-server model (depending on availability), and the resulting embeddings are then used for search on your device.

Chat: the app sends your chat messages to the backend. If the request uses vision, the app includes resized image attachments prepared for the AI request. The backend forwards the request to AI providers accessed via OpenRouter and returns the generated response to the app.

Feedback: when you submit feedback, the backend emails your message (and optional images) to our support inbox using SMTP. Feedback requests are rate-limited per device.

We process this data to provide and improve the App, enforce quotas, prevent abuse, provide support, and comply with legal obligations.

3. Third parties and sharing

We do not sell your personal data. We share data with service providers only as needed to run Screenshot AI Tool:

OpenRouter / AI providers (OCR, captioning, embeddings, chat): the backend uses OpenRouter to access AI models. Model families currently configured for production may include Google, Anthropic, OpenAI, Cohere, Qwen, and similar providers. Exact models can change as we maintain quality, cost, and availability. Image content and prompts you send for indexing or chat are processed by those providers under their terms and privacy policies.

Infrastructure: if Redis is configured, job state and quota counters may be stored there to support scaling and batch indexing. Hosting and networking providers process connection metadata necessary to deliver the service.

Apple: purchases and subscription status are handled through Apple / StoreKit. We receive transaction identifiers needed to verify entitlements; we do not receive your payment-card details.

Email: in-app Feedback is sent via SMTP (Resend) from noreply@mail.screenshotai.org to our support inbox. Public contact address: support@screenshotai.org.

Website analytics (Cloudflare Web Analytics): we may load a lightweight Cloudflare beacon on screenshotai.org to measure traffic and page performance (page views, referrers, countries, Core Web Vitals). This is privacy-oriented analytics, not ad tracking.

4. International transfers

Our backend and AI providers may process data in countries other than where you live. Where required, we rely on appropriate safeguards and contractual terms offered by those providers. By using the App, you understand that your indexing and chat requests may be processed outside your country of residence.

5. Data retention

On your device: OCR text, captions, embeddings, screenshot metadata (including file paths), and chat history are stored locally in SQLite databases. You can clear the local index and remove indexed entries from the database using the app’s clear and delete controls.

On our backend (index batch jobs): for batch indexing, the backend stores the derived job state (OCR text, captions, embeddings, and job progress/results metadata) so the app can poll job status. Job state is retained for up to 1 hour (TTL-based garbage collection), then removed.

On our backend (quotas and usage): we store per-device usage counters and hashed image identifiers (SHA-256 file hashes) to enforce indexing/chat limits. When Redis is enabled, some daily counters expire automatically; index hash sets and totals may persist until removed, depending on configuration.

Logs and metrics: we collect HTTP request metrics (route and status), and we record OpenRouter usage totals (token counts and cost) for billing/quotas. Error logs may include a pseudonymous device identifier and request metadata; we do not include full screenshot content in these metrics.

Feedback and support email: retention depends on our email provider and ordinary support practices. We keep correspondence as long as needed to resolve your request and meet legal requirements.

6. Your choices and rights

You control which folder the app indexes by selecting a folder during onboarding and by changing/removing that source in the app's settings. You can also clear or delete locally indexed items via the app.

Depending on where you live, you may have rights to request access, correction, deletion, or restriction of personal data we hold about you, or to object to certain processing. To exercise these rights, contact support@screenshotai.org. We may need enough information to verify your request (for example, relating it to a known device_id or App Store purchase).

7. Children's privacy

Screenshot AI Tool is not directed to children under 13 (or the minimum age required in your jurisdiction). We do not knowingly collect personal data from children. If you believe a child has provided personal data, contact us so we can take appropriate action.

8. Security

We use HTTPS for API traffic, store the device identifier in the system Keychain on macOS, and limit backend retention of image-derived job payloads. No method of transmission or storage is completely secure; please choose carefully which folders you index.

9. Changes

We may update this Privacy Policy when the App, our services, or legal requirements change. We will publish the updated version at this URL and revise the “Last updated” date. Material changes may also be noted in the App or on the website.

10. Contact

Privacy questions: support@screenshotai.org. Support page: screenshotai.org/support.html.