How noll.to Works
Translate. Download. Gone.
noll.to is the stateless translation service built for sensitive documents. We process your files without remembering them. No storage, no history, no trace.
Overview
Most translation tools trade your privacy for convenience. They store your history, train AI models on your text, and keep logs of your files.
We don't.
Our architecture is strictly zero-retention. We use ephemeral storage that exists only as long as your translation job is active. Once you download your files—or after a short timeout—everything is wiped. Forever.
The Process
1. Upload & Sanitize
You upload your documents via a secure, encrypted TLS 1.3 connection. We immediately strip non-essential metadata to protect your identity. We support PDF, Office formats (DOCX, PPTX, XLSX), and plain text.
2. Isolate & Translate
Each job runs in an isolated context. We route your content to our enterprise-grade engine (hosted in Azure EU data centers), which processes the text without storing it for training. Your data never touches a public model.
3. Encrypt & Protect
While processing, your data is protected by AES-256 encryption at rest. We generate time-limited SAS tokens for access—no permanent links, no public buckets.
4. Deliver & Destroy
You download your files. 30 minutes later—or immediately upon your command—we hard-delete the source files, the translated results, and the job metadata. It’s like it never happened.
Security & Privacy
Stateless by Design
We built noll.to to have no memory. This isn't just a policy; it's our code.
| Data Type | Retention Policy |
|---|---|
| Document Content | Deleted immediately after job expiration (30m) |
| File Metadata | Sanitized on upload, deleted with content |
| Translation Logs | Never stored |
| User Identity | Optional (Account required only for high-volume) |
Data Sovereignty
Your data stays in the European Union. We use Azure's EU regions for all processing to ensure strict GDPR compliance and data sovereignty.
FAQ
Q: Do I need to create an account?
A: No. Privacy shouldn't require a login. You can translate documents anonymously. We only ask for an account if you need higher volume quotas.
Q: Are my documents used to train AI models?
A: Never. We explicitly opt out of data logging and model training with our translation providers. Your intellectual property remains yours.
Q: What happens if I forget to download my files?
A: For your security, they are deleted automatically after 30 minutes. We cannot recover them. This is a feature, not a bug.
Q: Which file formats do you support?
A: We handle complex layouts in PDF, DOCX, PPTX, XLSX, HTML, and Markdown, preserving your formatting, images, and fonts.
Credits
noll.to uses a simple, transparent credit system based on page count. One credit equals one page processed.
How Credits Are Calculated
Your credit usage is calculated as:
Input Pages + (Output Pages × Number of Languages)
| Component | Description |
|---|---|
| Input Pages | The total page count of all documents you upload |
| Output Pages | The total page count of all translated documents (usually equals input) |
| Number of Languages | How many target languages you select |
Example
You upload a 10-page document and translate it into 3 languages (German, French, Spanish):
- Input Pages: 10
- Output Pages: 10 × 3 = 30
- Total Credits: 10 + 30 = 40 pages
Why This Model?
We charge for both input and output because:
- Input processing — We analyze and prepare your documents for translation.
- Output generation — Each translated version requires full rendering and formatting.
This ensures you only pay for what you use, with no hidden fees or per-character pricing surprises.
Free Tier
Every user gets 50 free pages per month — no account required. Need more? Check our plans.