Privacy Policy
How KontextFlux handles uploaded images, edit instructions, technical data, and short-lived results.
Updated July 30, 2026The short version
KontextFlux is designed for one-off image edits without an account. We process the image and instruction needed to complete your request, protect the service from abuse, and delete short-lived files on a limited retention schedule.
Do not upload material that you do not have the right to process or information that you would not want sent to a third-party AI provider.
Information we process
- The image you upload and the structured edit instruction you submit.
- Prepared image details such as file type, dimensions, and size.
- A random anonymous browser identifier stored in a signed, HttpOnly cookie.
- Privacy-preserving hashes derived from network and browser characteristics for rate limiting and abuse prevention. We do not intentionally store raw IP addresses in the application database.
- Job status, provider/model routing, error category, timing, and download events needed to operate and secure the service.
- Cloudflare may process ordinary edge, security, and Turnstile telemetry under its own terms and policies.
Why we process it
- To validate, queue, perform, store, and deliver the requested edit.
- To enforce individual, shared-network, burst, and global free-capacity limits.
- To detect unsafe or prohibited requests and investigate service failures.
- To maintain security, reliability, and aggregate product performance without analyzing your image for advertising profiles.
AI providers
KontextFlux sends the prepared image and edit instruction to a third-party AI inference provider. The provider may route the request to an underlying model provider. Provider handling, retention, and model-improvement practices can change under their applicable agreements and policies.
KontextFlux does not claim that uploaded images are never used for provider improvement or training unless that claim is supported by the active provider agreement. Avoid uploading highly sensitive or confidential material.
Retention
- Prepared input images are scheduled for deletion approximately one hour after upload.
- Successful output images are scheduled for deletion approximately 24 hours after completion.
- The normalized edit instruction is cleared from the job record after a terminal result or expiry.
- Minimal operational records, hashes, error categories, and aggregate capacity counters may remain longer for security, abuse prevention, accounting, and reliability analysis.
- Deletion is automated and may not be instantaneous during an outage or provider incident, but the service is not intended as permanent image storage.
Cookies and local browser data
The service uses a signed anonymous-session cookie to enforce free usage and protect private job results. The editor may also use browser memory or local storage for non-sensitive interface preferences. We do not use an account cookie or cross-site advertising cookie at launch.
Sharing
We share data with infrastructure and inference vendors only as needed to operate the service, comply with law, prevent harm, or protect legal rights. We do not sell uploaded images or edit instructions.
Your choices
- Close or reset the editor to discard the browser's local preview.
- Do not submit the AI job; local resize, compression, conversion, and prompt-building tools can run in your browser.
- Contact hello@kontextflux.com about a privacy request and include the job ID if it is still visible. We may be unable to identify anonymous use after short-lived records expire.
Changes
We may update this policy when providers, features, or legal requirements change. The updated date at the top identifies the current version.