September 12, 2025

• quay.io/deepgram/self-hosted-api:release-250912

  • Equivalent image to:
    • quay.io/deepgram/self-hosted-api:1.159.2

• quay.io/deepgram/self-hosted-engine:release-250912

  • Equivalent image to:

    • quay.io/deepgram/self-hosted-engine:3.101.1

  • Minimum required NVIDIA driver version: >=570.172.08

• quay.io/deepgram/self-hosted-license-proxy:release-250912

  • Equivalent image to:
    • quay.io/deepgram/self-hosted-license-proxy:1.9.1

• quay.io/deepgram/self-hosted-billing:release-250912

  • Equivalent image to:
    • quay.io/deepgram/self-hosted-billing:1.12.1

⚠️ Please be advised that this release is known to contain an Aura-2 TTS defect where the wrong conditioning priors are applied, resulting in the insertion of a random utterance. While the issue is rare, we strongly advise all Aura-2 TTS customers to update to release-250929 as soon as possible, where this behavior was resolved by a code change.

• GA support for Nova-3 monolingual German, Dutch, Swedish, Danish, Spanish, French, and Portuguese languages. Ask your Deepgram account representative for the new models.
• Improves Utterance End settings to allow lower minimum utterance_end_ms values, aligning with configured step size for self-hosted deployments.
• General Improvements — Keeps our software up-to-date.

Important: The onprem-* image tags are now deprecated and are no longer supported or updated. As announced in our previous release, we have transitioned to using only self-hosted-* repositories. Please update your deployments to use the self-hosted-* image tags listed above.