Trust
Verify the privacy promise.
OpenLLM Studio is built to keep your data where it belongs — on your machine. There is no telemetry, no analytics, and no outbound inference calls. If you want proof, the source is on GitHub.
Security principles
Zero telemetry
No crash reporters, no usage analytics, and no phone-home health checks. The app does not collect machine IDs or IP addresses.
Zero outbound calls
Inference, chat, and agent actions stay on your machine. After a model is downloaded you can disconnect from the internet entirely.
No account required
There is no cloud login, no API key gate, and no identity provider. Download and run.
Open-source by default
MIT-licensed source is on GitHub. Inspect the network layer, build from source, or audit every dependency.
Built for compliance
Designed to support SOC 2 evidence packages, HIPAA-ready access controls, GDPR data-residency snapshots, and immutable audit logs.
Local-first architecture
Models, embeddings, conversation history, and cache live on disk you control. Nothing is synced to a vendor cloud.
How to verify zero outbound calls
- Clone the repo from GitHub.
- Search for outbound URLs or telemetry endpoints.
- Run the app inside a network sandbox or firewall rule.
- Download a model, then disconnect from the internet.
- Confirm chat, agent, and editor continue to work.
Network proof at a glance
Need a compliance review?
Our enterprise team can walk you through architecture, audit logs, and air-gapped deployment options.