Releasesby GlyphVerify Callback Requests with Basic Auth or Extra Metadata
The dg-token header is no longer supported for verifying callback requests. Users must switch to Basic Auth or use a secret passed via Extra Metadata.
Deepgram Release notes from Deepgram
229 releases · 0 this month
Nova-3 Multilingual Expands to 55 Languages and Adds Language Hinting
Deepgram's new Nova-3 multilingual model supports 55 languages, code-switching, and language hinting via the /v1/listen API endpoints.
This Release Contains The Following Changes
- New Flux STT Runtime (action required) , Engine now runs Flux STT on a new runtime that loads only the .dgv2 Flux model files, for both flux-general-en and flux-general-multi . The new runtime does not use previously delivered .dg Flux model files: with only a .dg file present and [flux] enabled, Engine exits at startup with Can't find Flux model .
Update Nova-3 Keyterms Mid-Stream on Every Regional Endpoint
Mid-stream keyterm updates with Configure on /v1/listen now work on the EU ( api.eu.deepgram.com ), Australia ( api.au.deepgram.com ), and India ( api.in.deepgram.com ) regional endpoints , as well as the global endpoint ( api.deepgram.com ). They behave the same on every endpoint.
Update Nova-3 Keyterms Mid-Stream from the SDKs, Play Audio from the CLI, and Deepgram Skills v1.7.0
Deepgram Python SDK v7.13.0 , JavaScript SDK v5.15.0 , and Java SDK v0.11.1 let you reconfigure an active Nova-3 stream without reconnecting, including keyterms and numerals, and add custom Think-provider messages for Voice Agent sessions.
Correction: over-limit Configure keyterm updates on /v1/listen
The October 2 entry on mid-stream Nova-3 keyterm updates said that a Configure message whose keyterms exceed the 500-token limit returns an Error and leaves the stream's keyterms unchanged. That is not the current behavior.
Improved models
- Batch: nova-3-medical , English ( en , en-US , en-AU , en-CA , en-GB , en-IE , en-IN , en-NZ ) Streaming: nova-3-medical , English ( en , en-US , en-AU , en-CA , en-GB , en-IE , en-IN , en-NZ ) You do not need to change your API requests to access the improvements; the updates are live for all users. Available for hosted and self-hosted* customers.
Nova-3 Medical Multilingual: Dutch, English, French, German, Hindi, Italian, Japanese, Portuguese, Russian, and Spanish
The nova-3-medical model now supports multilingual transcription with language=multi . Nova-3 Medical can now transcribe medical audio in which speakers switch between Dutch, English, French, German, Hindi, Italian, Japanese, Portuguese, Russian, and Spanish. Available for both batch and streaming.
Improved models
Batch models Vietnamese ( vi ) Catalan ( ca ) Streaming models French ( fr , fr-CA ) Indonesian ( id ) You do not need to change your API requests to access the improvements to existing languages; the updates are live for all users. For more details and to see the full list of supported languages, visit the Models & Languages Overview page.
Update Nova-3 Keyterms Mid-Stream
Swap in the vocabulary for each step of a call without reconnecting. To replace a Nova-3 stream's keyterms during a /v1/listen session, send a keyterms array in a Configure message: { "type": "Configure", "keyterms": ["Deepgram", "customer service"] } Each keyterms array replaces the whole list, including keyterms set with the keyterm query parameter, and...
This Release Contains The Following Changes
- Flux TTS Watermarking (action required) , Flux TTS audio is now watermarked. Flux TTS deployments must add the watermarker model file ( watermarker. .dgv2 , provisioned alongside your Flux TTS models) and set watermarker_uuid under [flux_tts] in your Engine configuration. Engine will not start with Flux TTS enabled if the watermarker is missing.
Flux TTS Inline Controls: Pause and Pronunciation
Flux TTS ( /v2/speak ) now supports inline pause and pronunciation controls in the text you send, alongside speed . Pause control. Insert silence with an escaped marker such as \{pause:1s\} : 500–3000 ms in 100 ms steps, up to 8 per request. Available on batch (REST) requests. Pronunciation control (Early Access).
Toggle Numerals Mid-Stream on Flux STT
Switch Flux STT to digits for a PIN, phone number, or order number, then back to words, without reconnecting. To turn Numerals on or off during a Flux STT stream, send numerals as a boolean in a Configure message: { "type": "Configure", "numerals": true } The update applies to transcripts Flux STT sends after it processes the message, and ConfigureSuccess...
Improved models
Batch and streaming models German (Switzerland) ( de-CH ) Portuguese ( pt , pt-BR , pt-PT ) Lithuanian ( lt ) Streaming models Flemish ( nl-BE ) You do not need to change your API requests to access the improvements to existing languages; the updates are live for all users.
Improved models
Batch and streaming models Polish ( pl ) Danish ( da ) Batch models Vietnamese ( vi ) Flemish ( nl-BE ) Streaming models Estonian ( et ) Macedonian ( mk ) Urdu ( ur ) Italian ( it ) Lithuanian ( lt ) You do not need to change your API requests to access the improvements to existing languages; the updates are live for all users.
Correction: Browser Agent SDK has no client-side VAD
The Browser Agent SDK never shipped client-side Silero voice activity detection. Documentation published with the SDK described it as an available feature, and that documentation was wrong. The following did not exist in any released package: The vad option on MicrophoneOptions in @deepgram/agents . The speechThreshold and silenceThreshold VAD options.
Nova-3 Pharma: Speech-to-Text for Pharmaceutical Use Cases (English)
We have released nova-3-pharma , a new Nova-3 model purpose-built for pharmaceutical vocabulary, with a focus on accurate drug-name recognition. It is designed for pharmacy and healthcare voice-agent workflows where getting the medication right matters most. Available in English for both batch and streaming.
This Release Contains The Following Changes
- NVIDIA Blackwell GPU Support (GA) , Blackwell-generation GPUs are now generally available for self-hosted deployments. See Drivers and Containerization Platforms for driver requirements and Model and GPU Compatibility for which models run on which GPUs.
Java SDK v0.10.0
Deepgram Java SDK v0.10.0 adds sendForceEndTurn(...) , Agent Flux TTS expressivity , and the expanded Flux TTS speed range. This pre-1.0 release includes breaking changes: it removes SpeakV2Speed and its named speed constants, plus the unsupported AURA2PERSEO_IT model constants and visitor methods.
@deepgram/react 0.2.0
@deepgram/react 0.2.0 adds runtime Voice Agent controls and makes React session lifecycle handling safer. The release updates the React layer for @deepgram/agents 0.1.2 and @deepgram/sdk 5.9.0.
FunctionCallCancelled
Clients now receive an explicit signal when a client-side function call they were sent is cancelled because the user started speaking again, either inside the speculative window (the turn resumed) or after the turn was confirmed (barge-in). Only calls the client already received are announced.
Improved models
Batch and streaming models Polish ( pl ) Hebrew ( he ) Batch models Macedonian ( mk ) Streaming models Estonian ( et ) Latvian ( lv ) Lithuanian ( lt ) Malay ( ms ) You do not need to change your API requests to access the improvements to existing languages; the updates are live for all users.
Wider speed range for Flux TTS
Flux TTS speed now runs from 0.5 to 1.5 in 0.05 increments, up from 0.85 – 1.15 . The default stays 1.0 , and every previously accepted value still works. The wider range applies everywhere speed is accepted on Flux TTS: the /v2/speak streaming WebSocket (at connection and mid-session with Configure ), the /v2/speak batch REST endpoint, and...
Flexible turn-taking control for Flux
Flux now gives you full control over how turns end. Beyond Flux's native end-of-turn detection, you can override it, suppress it entirely, or blend the two , and switch between these modes on the fly, even for a single turn, with a Configure message. Turn-taking goes from a fixed behavior to something you shape around each moment of a conversation.
Nova-3 Adds Assamese, Mongolian, and Pashto, Plus Improved Models for Czech, Danish, Swedish, Tagalog, and Turkish
We have added Assamese, Mongolian, and Pashto as new Nova-3 languages and released improved Nova-3 monolingual models for several existing languages. These updates expand language coverage and enhance transcription quality for batch and streaming workloads.
eot_threshold maximum raised to 1.0
agent.listen.provider.eot_threshold now accepts values up to 1.0 . Set it to 1.0 to fully suppress natural end-of-turn detection and own turn taking with ForceEndTurn . For details, see Configure the Voice Agent and Bring Your Own Turn Detection .
Deepgram CLI 0.3.0: Flux TTS and Flux STT
Deepgram CLI 0.3.0 is now available. It completes CLI support for Flux TTS , which reached general availability on 12 August, fixes Flux STT streaming, and enforces the CLI's exit-code contract. Upgrading from 0.2.26 picks up everything in this release. 0.2.27 was tagged but never published, so its changes arrive here.
Improved models
Batch and streaming models Russian ( ru ) Slovak ( sk ) Slovenian ( sl ) Urdu ( ur ) Streaming models Hungarian ( hu ) Macedonian ( mk ) You do not need to change your API requests to access the improvements to existing languages; the updates are live for all users.
Numerals Support Now Available for 4 New Languages: Bulgarian, Chinese (Cantonese, Traditional), Malay, and Korean (Monolingual Models)
Supported languages and language codes: Bulgarian ( bg ) Chinese (Cantonese, Traditional) ( zh-HK ) Malay ( ms ) Korean ( ko , ko-KR ) You can now use Deepgram’s Numerals feature with monolingual models for Bulgarian, Chinese (Cantonese, Traditional), Malay, and Korean.
This Release Contains The Following Changes
- Flux TTS (Early Access) , Flux TTS is now available in self-hosted deployments on the new /v2/speak endpoint. See Deploying Flux TTS for setup, and the section above for details. Nepali and Punjabi Support , Nova-3 monolingual models now support Nepali ( ne ) and Punjabi ( pa , pa-IN ), in both batch and streaming. See the full announcement for details.
Nova-3 Adds Armenian, Plus Improved Models for Tamil, Indonesian, and Belarusian
We have added Armenian as a new Nova-3 language and released improved Nova-3 monolingual models for several existing languages. These updates expand language coverage and enhance transcription quality for both batch and streaming workloads.
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.
Nova-3 Model Update
🌏 Nova-3 monolingual models now support Punjabi and Nepali with the following language codes: Punjabi: pa , pa-IN Nepali: ne Batch and streaming models are both available for these languages. Access this model by setting model="nova-3" and the relevant language code in your request.
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.
Fixed
Resolved a sample rate issue that prevented some users from using the Flux TTS early access model in the Voice Agent.
FIPS 140-3 Images (General Availability)
FIPS 140-3 images are now generally available for all four self-hosted services, published under the -fips tag suffix: quay.io/deepgram/self-hosted-api:release-260728-fips quay.io/deepgram/self-hosted-engine:release-260728-fips quay.io/deepgram/self-hosted-license-proxy:release-260728-fips quay.io/deepgram/self-hosted-billing:release-260728-fips These...
Java SDK v0.7.0
Deepgram Java SDK v0.7.0 is now available. This release adds the numerals parameter to Listen V2 WebSocket connections via V2ConnectOptions and ships new Aura-2 multilingual TTS voices. This release includes a breaking change. For the full release notes, see deepgram-java-sdk v0.7.0 .
New Models Released for Multiple Languages
We have released new Nova-3 monolingual models for the following languages. These additions expand language coverage for both batch and streaming workloads.
Enabling Numerals
Set numerals=true as a query parameter when you open the WebSocket connection: wss://api.deepgram.com/v2/listen?model=flux-general-en&numerals=true&encoding=linear16&sample_rate=16000 On Flux, numerals is set at connection time only and cannot be changed through the Configure message.
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.
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...
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.
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.
Deprecation: Claude Sonnet 4
claude-sonnet-4-20250514 is now deprecated. Migrate to claude-sonnet-5 or another supported Anthropic model. This model has been retired by Anthropic and will produce an error if specified in your settings configuration. For the full list of supported models and pricing tiers, see the Voice Agent LLM Models documentation .
This Release Contains The Following Changes
- Flux Word-Level Timestamps , each word in the words array now includes start and end times on the Flux streaming endpoint ( /v2/listen ). This is an additive change. If your client library rejects unrecognized response fields, update to a current SDK version , for the Deepgram Python SDK, use 6.0.1 or later, or 5.3.4 if you are on the 5.x line.
🆕 Redaction: Expanded Language Coverage & Flux Support
- Number redaction ( numbers , true , aggressive_numbers ) now supports these languages on the Hosted API: Danish, Dutch, English, French, German, German (Swiss), Italian, Norwegian, Polish, Portuguese, Spanish, and Swedish. On multilingual models, number redaction applies the correct rules to each detected language within a transcript.
Australia Endpoint Now Generally Available
The Deepgram Australia endpoint ( api.au.deepgram.com ) is now generally available for customers requiring data processing within Australia.
UpdateListen: On-the-Fly Listen Configuration
You can now update the Listen configuration during a conversation without restarting the session. Send an UpdateListen message to tune: eot_threshold , end-of-turn confidence threshold eager_eot_threshold , eager end-of-turn confidence threshold eot_timeout_ms , hard timeout for turn completion keyterms , keyword boost list language_hints , language bias...
Action Required: Engine Container GPU Environment Variables
The Engine container change previewed in the May 28, 2026 release has shipped in this release. The Engine container now requires two environment variables to access the GPU: NVIDIA_VISIBLE_DEVICES=all NVIDIA_DRIVER_CAPABILITIES=compute,utility If they are not set, the Engine container will fail to start after you upgrade to release-260611 .
🆕 Improved Nova-3 Medical Batch Model Released
We’ve released an upgraded Nova-3 Medical batch model with improved medical term recognition. Key Improvements: Expands the model’s medical vocabulary with new medical terms Boosts medical term recognition to 97.20% KRR Maintains WER parity with no observed regression on common or conversational terms
Preparation for a future Engine container change
The official Helm chart ( 0.37.0 and later) and the Docker and Podman compose files in deepgram/self-hosted-resources now set NVIDIA_VISIBLE_DEVICES=all and NVIDIA_DRIVER_CAPABILITIES=compute,utility on the Engine container.
Deprecation Notice
The Gemini 2.5 Flash family of models is deprecating in October. Start testing newer models now to ensure a smooth migration. For more details on Gemini model deprecations, see Google's Gemini deprecations page . For the full list of supported models and pricing tiers, see the Voice Agent LLM Models documentation .
🆕 Profanity Filtering for Multilingual Models
Deepgram’s Profanity Filtering feature is now available for all multilingual models: Nova-2 multilingual , Nova-3 multilingual , and Flux multilingual ( language=multi ). You can enable profanity filtering in your API requests by setting the profanity_filter=true parameter.
Gemini 3.1 Flash Lite Now Available
gemini-3.1-flash-lite is now available as a managed Google LLM in the Voice Agent API. This Standard tier model replaces the preview version. Set the model in your agent configuration: { "agent": { "think": { "provider": { "type": "google", "model": "gemini-3.1-flash-lite", "temperature": 0.5 } } } }
Numerals Support Now Available for 3 New Languages: Russian, Romanian, and Hebrew (Monolingual Models)
Supported languages and language codes: Russian ( ru ) Romanian ( ro ) Hebrew ( he ) You can now use Deepgram’s Numerals feature with monolingual models for Russian, Romanian, and Hebrew. Numerals converts spoken numbers into digits (for example, "three hundred" → "300") in your transcript, helping you create more accurate and easily processed results.
How to Use Profanity Filtering
To enable profanity filtering, add the profanity_filter=true parameter to your Deepgram API request: For more details, supported languages, and additional options, visit the Profanity Filter page .
Diarization v2: Improved Batch Speaker Diarization
A new batch diarization model is available today via the diarize_model API parameter. Deepgram is rolling out v2 of our batch speaker diarization model. v2 is a new architecture available today on an opt-in basis through the new diarize_model parameter.
Nova-3 Multilingual Model Update
Numerals Support Expanded for Nova-3 Multilingual Numeral formatting is now supported for all Nova-3 multilingual languages , except Hindi and Japanese. This enhancement means Nova-3 multilingual can now convert spoken numbers to digits (e.g., "three hundred" → "300") for English, Spanish, French, German, Russian, Portuguese, Italian, and Dutch.
Browser Agent SDK
The Browser Agent SDK is now available , four composable packages that connect any web app to the Voice Agent API: @deepgram/agents-widget , drop-in widget with six layouts (sidebar, floating, inline, button, embedded, or orb). No framework required.
Aura-2 Voice Controls , Speed and Pronunciation
Aura-2 TTS voices now support runtime speed and pronunciation controls in English and Spanish, available on both batch and streaming WebSocket endpoints. Together, these controls give developers finer-grained tools to improve naturalness , tuning pacing and correcting pronunciation to match their needs.
About and activity
Speech models and the voice agent API.
Releases per month
21May
11Jun
26Jul
38Aug
25Sep
22Oct
Official release-notes pages 5
| Release-notes page | Site | Feed |
|---|---|---|
| Changelog developers.deepgram.com | developers.deepgram.com | None |
| December 12, 2025 developers.deepgram.com | developers.deepgram.com | None |
| July 9, 2026 developers.deepgram.com | developers.deepgram.com | None |
| June 30, 2026 developers.deepgram.com | developers.deepgram.com | None |
| May 4, 2026 developers.deepgram.com | developers.deepgram.com | None |