Releasesby GlyphLatency reporting
LatencyReport provides a detailed LLM, TTS, and end-to-end latency breakdown for each turn, including ttt_token_latency , ttt_text_latency , ttt_tool_latency , ttt_thinking_latency , tts_latency , and total_latency .
Also shipped Jul 9, 2026Deepgram in July 2026 →
Change to default pipeline editor for workspaces with the compliance security profile enabled
The Lakeflow Pipelines Editor has been generally available since May 2026. In August, it will become the default editor for developing and debugging pipelines in workspaces with the compliance security profile enabled. At that time the legacy editor will be removed.
Strac connector (Beta)
The Strac connector is now available in Lakeflow Connect in Beta . Workspace admins can control access to this feature from the Previews page. The connector ingests data-security audit events from your Strac instance into Databricks. Authentication uses a Strac API key. See Strac connector .
Protobuf tensor input for custom model serving endpoints (Public Preview)
For custom model serving endpoints with a tensor-based input signature, you can now send the request body as a serialized KServe v2 ModelInferRequest with Content-Type: application/x-protobuf , instead of JSON. This feature is in Public Preview and is only available on endpoints deployed after July 9, 2026. See Protobuf tensor input .
Genie One enhancements
- Document citations : Documents now show citations that link content back to the data and sources Genie One used to generate the document. See View citations . Document version history : Documents now keep a version history so you can review earlier versions. See View version history .
Front-end Private Link for account-level Genie One (Beta)
You can now configure front-end Private Link for account-level Genie One. Turn on the Front-end Private Link for Custom URLs and Account preview to get started. See Configure inbound Private Link for account-level resources .
Dashboard enhancements and bug fixes
- Import BI files is in Public Preview : Importing Power BI and Tableau reports as AI/BI dashboards with Genie Code is in Public Preview. You can also start an import directly from the dashboard list page. See Import BI files using Genie Code .
More from DeepgramEvery Deepgram release →
InjectAgentMessage: behavior field and new interrupt mode
The InjectAgentMessage message now accepts a behavior field that controls how an injected message interacts with any in-progress user or agent turn. The default and queue behaviors were previously documented; the new interrupt behavior is introduced with this release. default , the agent speaks only if neither the user nor the agent is mid-turn.
Container Images (release 260714)
- quay.io/deepgram/self-hosted-api:release-260714 Equivalent image to: quay.io/deepgram/self-hosted-api:1.192.5-1 quay.io/deepgram/self-hosted-engine:release-260714 Equivalent image to: quay.io/deepgram/self-hosted-engine:3.119.1-3 Minimum required NVIDIA driver version: >=570.172.08 quay.io/deepgram/self-hosted-license-proxy:release-260714 Equivalent image...
This Release Contains The Following Changes
- Number and Percentage Formatting , decimal values now keep their trailing zeros, so "zero point zero zero percent" renders as 0.00% (previously 0% ); and Malay transcripts now convert spoken numbers to numerals.
Improved Models Released for Multiple Languages
We have released improved Nova-3 monolingual models for the following languages. These updates enhance transcription quality and accuracy for both batch and streaming workloads.
🆕 Improved Models
Batch models Ukrainian ( uk ) Catalan ( ca ) Turkish ( tr ) Romanian ( ro ) Finnish ( fi ) Malay ( ms ) Greek ( el ) German ( de ) Dutch ( nl ) French ( fr , fr-CA ) Streaming models Catalan ( ca ) Turkish ( tr ) Malay ( ms ) You do not need to change your API requests to access these improvements; the updates are live for all users.
🆕 Word-Level Timestamps Now Available in Flux
Flux now includes word-level timestamps in its responses. Each word in the words array carries start and end times (type double ) alongside the existing word and confidence values.
SDK Compatibility
Most integrations will continue to work without changes. The exception is clients configured to reject unrecognized fields, which may return errors when the new fields appear. Here is what we recommend: Python : Versions 6.0.1 and later are unaffected. If you are on the 5.x line, update to 5.3.4.
🆕 Numeral Formatting for Flux
Flux now supports the numerals feature, which converts numbers from written format to numerical format (for example, "nine hundred" becomes "900"). Numerals are available on both Flux models: Flux English ( model=flux-general-en ) , full numeral formatting.