SUPERWISE Sentinel 1.40.0: An image sharpness guardrail and a clearer CLI

SUPERWISE® Sentinel 1.40.0 is live. It adds a second image guardrail, makes language detection more accurate, and lets the Sentinel CLI show a Sentinel’s full policy.

Version 1.40.0 was released on October 25, 2026. Here’s what’s in it.

Guardrails

• New guardrail: Image sharpness. Blocks a request when an image attached to the prompt is too blurry. Set a minimum sharpness score: an image that scores below it is rejected with your error message and never reaches the model. A blurry image wastes tokens and weakens a vision model’s answer. It’s Sentinel’s second image guardrail, after Image resolution.
• More accurate language detection. The Language check guardrail, added in 1.39.0, now identifies a prompt’s language more accurately.

See the docs for Image sharpness, Language check and testing a guardrail before you publish.

Sentinel CLI

• See a Sentinel’s full policy. sentinel gateway status --details prints the Sentinel’s guardrails and token optimization settings alongside its health and version.
• Clearer errors. The CLI checks up front that you have an Admin role and that your --sentinel-id exists, and tells you exactly what’s wrong instead of failing with an unclear error.
• Smoother updates. After a gateway update, the CLI prompts you to restart a running gateway. When the health check can’t be reached, it falls back to a test ping.

Also in this release

Clearer endpoint descriptions in the API reference, and bug fixes.

Get started

New to Sentinel? See what it does to a real prompt in 30 seconds, with no install and no sign-up, or get Sentinel for Windows. Every plan is free for 30 days.

Read the full 1.40.0 release notes and the Sentinel documentation.