The situation
You take over someone else's code. It is hard to read. The code contains company names, addresses and credentials.
How it works
You paste the code.
The Schleuse replaces secrets
Names, addresses, identifiers and credentials are replaced with stand-in words. This happens on your side, not in the cloud. Amounts and dates stay as they are.
The AI answers, the Schleuse translates back
The answer contains the stand-in words. On your side they turn back into the real names.
Example
Explain what this code does. Suggest improvements.
# fetch invoices for Müller Landtechnik GmbH url = "https://api.mueller-landtechnik.intern/v1" token = "tk-live-81d9f0aa27"
# fetch invoices for Company Alpha url = "https://Server_A/v1" token = "Token_A"
This example is made up.
What you get
You get an explanation and suggestions.
Good to know
Credentials do not belong in code. The Schleuse protects them, but it does not replace proper storage.
Zone of this use case: Yellow
Only through the Schleuse. And only to targets you have approved.
Status of the Schleuse
Version 0.1 Alpha. So far tested only with made-up examples. Not reviewed by third parties. No legal advice.
Related use cases
Analyse an error log
The AI finds the cause. Passwords and customer names stay secret.
Read more →Software development and ITReview a configuration file
The AI looks for errors and insecure settings.
Read more →Software development and ITUse database extracts for tests
The AI gets sample data, but no real people.
Read more →Would you like to try this?
We are looking for pilot partners. We test the Schleuse with made-up or cleared examples first. Only then does real text go through.
This page is also available as Markdown for AI reading systems. /en/use-cases/explain-and-review-code.md