Upload & Validation
Files stream directly from your browser to our EU-based servers over encrypted connections. We validate file size, format, and integrity before accepting the upload. If any check fails, the upload is immediately rejected with a clear error message.
Batch uploads are subject to your tier's limits—we cap both file count and total size to ensure fair resource distribution.
Credit Debit & Queueing
Once validated, each file becomes a job. We verify you have sufficient credits (or daily allowance for free tier users) before queueing. Credits are debited when a worker starts processing—not when the job enters the queue. If the job fails at any point after debiting, credits are automatically refunded.
Jobs enter a queue managed by our fair scheduler. Paid users get priority (Business, Pro, then Free), but no single user can monopolise workers. Within each tier, jobs are distributed fairly—not first-come-first-served.
Processing
A worker picks up your job and compresses it using our optimised pipeline. Processing includes:
- EXIF stripping by default — Removes location data, camera settings, and other metadata for privacy and smaller files
- Optimised quality presets — Balanced for excellent visual quality and maximum size reduction
- Format-specific algorithms — Each format uses industry-leading compression techniques optimized for that format
- Transparency handling — PNG and WebP preserve transparency; no unwanted background colors
Typical compression times: 0.1-2 seconds for images depending on resolution. PDFs take longer based on page count and content complexity.
Storage & Auto-Delete
Outputs are encrypted at rest. All files—input and output—are automatically deleted after 1 hour on all tiers, by design. We do not create backups. Once deleted, files are unrecoverable.
We never use your content to train AI models. See our retention policy for the full rationale.
Related Topics
Why is compression so fast?
Learn about our architecture and optimization techniques
How do you encrypt my files?
Technical details about TLS 1.3 and AES-256 encryption
What metadata do you keep?
What we log about jobs—and what we never store
Can I process files in bulk?
Batch limits and queue capacity by tier