How your files are handled
Data_Link is built so that the file you upload stops existing as soon as the work is done. This page describes exactly what happens and what is kept.
The lifecycle of a file
- Upload
The file is streamed to object storage. It is never written to a database and never read into memory in full.
- Processing
A worker downloads it, reads it in blocks, and writes the result to a separate file. No copies are made anywhere else.
- Source deleted
When the job finishes, the original file is deleted from storage. It is part of the job, not a scheduled cleanup that might not run.
- Result expires
The cleaned file stays available for 24 hours so you can download it, then it is deleted automatically. On the free plan, one hour.
What is kept after that
Kept
- Your email address, for your account
- The file name and its size
- How many records were processed, removed and kept
- When the job ran and how long it took
Not kept
- The contents of your file
- Any individual record, value or column
- The source file, once the job completes
- The result, past its expiry window
The statistics on your dashboard are counters, not samples. They are produced while the file streams past and describe quantities, never content.
API keys
Your key is the credential for every request. It is stored in your browser only, which is why the console offers “Forget this device” — use it on shared machines. If a key is exposed, regenerate it from the console and the old one stops working.
What we do not claim
Data_Link is a data cleaning service, not a compliance product. We do not hold a security certification, and nothing here should be read as legal advice about the regulations that apply to your data. What we describe on this page is how the system behaves — you decide whether that fits your obligations.
If you handle personal data and need masking or anonymisation rather than deduplication, that is Data_Link Transform, which is still in development.