How sven protects your account, your data and your money, what to expect when you connect a computer, and how to report a problem.
Version 2026-10-29
Last updated: 29 October 2026.
Security and privacy are designed in from the start, not added afterwards. We check our production systems continuously with automated security audits. When an audit finds something, we fix it in the configuration that builds our systems, so the fix cannot be undone by accident.
Signing in is handled by an established identity provider. Every request to sven is checked to see who is asking and what they may do.
Every customer's data is kept separate from every other customer's. We enforce this twice, once in the application and again in the database, so a mistake in one place cannot expose another customer's data.
A key is shown to you once, when you create it. We store only a one-way fingerprint of it, so not even we can read it back. Keys expire, and you can revoke one at any time.
Spending limits on a key and a monthly budget for your workspace stop costs before the work starts: a call that would go over is refused rather than run and billed afterwards.
Payments are handled by Stripe. Your card details go to Stripe and never reach our servers.
All traffic to sven is encrypted: the site and the API are reachable over HTTPS only, and your browser is told to insist on it. Secrets for the services you connect are encrypted at rest.
Administrative access to our production systems requires a cryptographic key; passwords are never accepted, and changes are recorded in an audit trail.
You can connect your own computer with the sven agent (sven rc), so that sven can work on your files. This is what protects you when you do, with sven 4.2.2 or later.
Commands the agent runs in a terminal run with the permissions of the user who started it. That is why we recommend running the agent in our hardened container setup, which adds a boundary around the project. See the sven documentation for how to set it up.
When sven runs an agent for you in a web conversation, that conversation's agent runs isolated from the others and from the systems that hold account and payment data. It can only read public web pages and search the web.
Only what must be public is reachable from the internet. Security updates install automatically, and the servers restart to apply them. Every service runs with the least privileges it needs. Logs are kept for a limited time.
The database is backed up daily, and backups are kept for two weeks. They are accessible only to the operators of the service.
How long each kind of data is kept is set out in the Privacy Policy.
If a security incident affects your data, we will inform affected customers without undue delay.
If you think you have found a security problem, write to info@swedishembedded.com. We read every report. Please do not disclose an issue publicly until we have fixed it.