Containers
Your container, inside a TEE.
Run any Docker image inside a trusted execution environment (TEE), measured for anyone to verify. Clients verify its remote attestation before they send a request.
Join waitlistFeatures
Bring the image you already have.
No rewrite and no special build. The image runs as you built it.
Any Docker image
Pin your image by its SHA-256 digest, set its ports, secrets, and resources, and it runs inside an enclave.
Remote attestation
Clients verify the enclave's remote attestation before any request reaches your workload.
Measured and logged
Each release is measured and recorded in a transparency log, so anyone can match what runs to its source.
Updates with no downtime
Tag a new release, and the enclave moves to it while the service keeps running.
Private registries
Pull images from a private registry. The verification stays pinned to the digest of the image.
Confidential GPUs
Add one GPU or several: NVIDIA Hopper and Blackwell in confidential computing mode, dedicated and never preempted.
Platforms
The TEEs of AMD, Intel, and NVIDIA.
- CPU
- AMD SEV-SNP · Intel TDX
- GPU
- NVIDIA Hopper · Blackwell
- GPU mode
- Confidential computing
- GPUs
- One or more
- Measurements
- Transparency log
- Verification
- Remote attestation
How it works
From image to enclave in four steps.
- 01
Pin the image
Name your image by its SHA-256 digest, with its ports, secrets, and resources, in one config file.
- 02
Tag a release
Each tag is measured and recorded in a transparency log before it deploys.
- 03
Run it in a TEE
The image boots inside an enclave on AMD SEV-SNP or Intel TDX, with any GPU in confidential computing mode.
- 04
Verify, then connect
Clients verify the enclave's remote attestation against the logged measurement, then send their first request.