Our Commitment
We will never use your files to train AI models. This is a core architectural decision, not a policy that can change. Your files exist on our servers only long enough to be converted and downloaded.
No AI Training, Period
Your uploaded files are never used to train AI models, machine learning systems, large language models (LLMs), image generators, music generators, voice synthesis systems, or any automated learning system. This applies to:
- Our own internal systems
- Third-party AI services
- Research partnerships
- Any future AI initiatives
We don't analyze file contents for patterns, extract features for learning, or aggregate data for model improvement. Conversion is conversion, nothing more.
No Third-Party Data Sales
We never sell, license, or share your file contents with third parties. Your data is not a product. We make money from subscriptions, not from monetizing your files.
This includes:
- No selling to data brokers
- No sharing with advertisers
- No partnerships that involve access to your files
- No "anonymized" dataset licensing
What We Actually Keep
We retain only minimal, aggregate operational metrics:
- Job counts — How many conversions were processed
- Processing times — How long conversions took (for performance optimization)
- File sizes — Input/output sizes (for capacity planning)
- Format types — Which conversion types are popular
- Error rates — To improve reliability
We explicitly do not keep: file contents, filenames, EXIF metadata, document text, image pixels, or any data that could identify what was in your files. See What metadata do you keep? for complete details.
Technical Safeguards
Our architecture enforces this policy technically, not just procedurally:
- Auto-delete — Files are automatically deleted within 1 hour of conversion
- No backups of files — We don't create backups that could be used later
- Encryption — Files are encrypted in transit (TLS 1.3) and at rest (AES-256)
- Isolated processing — Each job runs in isolation with no shared state
- EU infrastructure — Processing happens on EU-based servers under GDPR jurisdiction
Why This Matters in 2025
Recent years have seen major companies quietly using customer data to train AI models, often buried in terms of service updates. We believe this is wrong.
When you use a file converter, you expect one thing: file conversion. Not to become training data for someone else's AI product. We respect that expectation completely.
What About AI-Powered Features?
If we ever add AI-powered features (like smart cropping or content-aware processing), they would use pre-trained models only. Your files would be processed by the model, never used to improve it. We would also make such features clearly opt-in with explicit disclosure.
GDPR Alignment
Our no-training policy supports GDPR principles:
- Purpose limitation — Data used only for the stated purpose (conversion)
- Data minimization — We keep only what's necessary for operations
- Storage limitation — Auto-delete after 1 hour
- Privacy by design — Architecture prevents misuse by default
The Bottom Line
Your files are your files. We convert them, you download them, and then they're gone. No training, no selling, no exceptions. This is fundamental to how we built Convert.FAST and will never change.