Your data never trains a model
3 months ago
Model providers don't train on qlane's AI requests, under their terms or our AI gateway provider’s agreements; zero data retention is a setting for each organization.
Model providers never use qlane's requests for training, under their terms or our AI gateway provider’s agreements with them. That holds for every call, from planning tests to chat.
Zero data retention, where the provider keeps no prompt or output after the request (apart from up to 5 minutes of prompt caching), is a setting for each organization. Since 12 October 2026 it is off unless it has been turned on for your organization; until then, it applied to every request.
The new security page at qlane.ai/security describes how qlane handles your data and lists the sub-processors involved.
- No training by model providers on any AI request, under their terms or our AI gateway provider’s agreements with them; zero data retention for organizations that have it turned on.
- A security page at qlane.ai/security, with an updated list of sub-processors.
- Legal pages now match the published security posture.
- Sandboxes get a public address (QLANE_SANDBOX_PUBLIC_URL) on qlane's own domain, so SSO and OAuth redirects work inside them.