Deployment options
Self-host for full control, or let us handle operations. You choose the responsibility boundary.
Self-hosted
Deploy on your infrastructure under AGPLv3, or with commercial licenseYou can run Caution internally without AGPL obligations via commercial licensing. Contact us at info@caution.co to obtain the commercial license..
Includes:
- Full control over deployment and infrastructure
- Community support
- You manage operations and uptime
Managed on-prem
Your servers, our expertise. We handle updates, monitoring, and support.
Includes:
- Deployment on your own servers or cloud
- Caution-managed updates and operations
- Zero-downtime updates
- Priority support
Fully managed
We host and operate everything. Ship faster with minimal overhead.
Includes:
- Hosted in Caution's environment
- We handle infra, updates, and reliability
- No operational overhead
- Priority support
Included with every deployment:
- Fully open source codebase
- Verifiable builds and attestation
- Reproducible enclave images
- Hardware-backed isolation
- Transparent trust model
- Automated provisioning
One workflow, any TEE
Caution runs across AWS Nitro today, with Intel TDX, AMD SEV-SNP, and TPM 2.0 attestations coming in 2026.
Learn more
Frequently asked questions
Choose self-hosted if you want full control and have the capacity to manage infrastructure yourself. It is fully open source and licensed under AGPLv3. If you want to evaluate Caution internally without AGPL obligations, contact us at info@caution.co for commercial licensing.
Choose managed on-premises if you need workloads on your own servers but want Caution to handle updates and operations.
Choose fully managed if you want the fastest path to production with zero operational overhead.
Early access gives you a pre-release version of Caution that you can use either through the hosted alpha program or by self-hosting the open source stack. At this stage, Caution supports AWS Nitro enclaves only.
The software is under active development and may introduce breaking changes. Alpha is intended for teams that want to experiment with verifiable compute, test workflows end to end, and help shape the product. Learn more here.
No. The alpha is for evaluation and experimentation, not for high value production workloads. You can deploy real code and see the full verifiable compute workflow, but you should treat this as a testing environment.
Contact us at info@caution.co.