AssemblyAI Release Notes

Name: AssemblyAI
Brand: AssemblyAI

Follow AssemblyAI to add their release notes to your feed!

126 release notes curated from 1 source by the Releasebot Team. Last updated: Jul 15, 2026

Get this feed:

Jul 14, 2026
Date parsed from source:
Jul 14, 2026

First seen by Releasebot:
Jul 15, 2026
AssemblyAI

Introducing the Sync API: transcripts in a single call

AssemblyAI adds a Sync API for instant Universal-3.5 Pro transcripts in one HTTP request with no polling or WebSockets.

The Sync API is now available: send a short audio clip in one HTTP request and get a finished Universal-3.5 Pro transcript back in the same response, in ~134 ms at p50. No polling, no WebSocket, no job to manage…
Original source
Jul 13, 2026
Date parsed from source:
Jul 13, 2026

First seen by Releasebot:
Jul 15, 2026
AssemblyAI

Fix: Payment Method Updates in the Dashboard

AssemblyAI fixes a Stripe schema issue that restores payment method updates in the dashboard.

A Stripe response-schema change broke updating payment details, returning a 400 on the payment-method endpoint. This is now fixed, and you can update your payment method in the dashboard again…
Original source
All of your release notes in one feed

Join Releasebot and get updates from AssemblyAI and hundreds of other software products.

Create account
Get updates with:
Jul 13, 2026
Date parsed from source:
Jul 13, 2026

First seen by Releasebot:
Jul 15, 2026
AssemblyAI

Expanded Voice Catalog for the Voice Agent API

AssemblyAI expands its Voice Agent API voice catalog with new in-house Pocket voices, now in production and the docs and playground.

The Voice Agent API voice catalog has expanded with new in-house (Pocket) voices, deployed to production and rolling into the docs and playground. Pick a voice_id from the voices reference…
Original source
Jul 11, 2026
Date parsed from source:
Jul 11, 2026

First seen by Releasebot:
Jul 15, 2026
AssemblyAI

Breaking: Async Default Model Changes to Universal-3.5 Pro on August 7

AssemblyAI updates speech model routing to use the default model and plans a switch to universal-3-5-pro in 2026.

Requests that omit speech_models (or pass a legacy speech_model name) route to the default model. The default is universal-2 today and switches to universal-3-5-pro on August 7, 2026. On the same date, requests pinned…
Original source
Jul 11, 2026
Date parsed from source:
Jul 11, 2026

First seen by Releasebot:
Jul 15, 2026
AssemblyAI

Onboarding Configuration Wizard in the Dashboard

AssemblyAI adds a configuration wizard for new accounts to guide product selection and setup from onboarding and the dashboard.

New accounts now get a configuration wizard that walks through product selection and setup. It is available from the onboarding page and the home-page call to action in the dashboard…
Original source
Similar to AssemblyAI with recent updates:
Jul 11, 2026
Date parsed from source:
Jul 11, 2026

First seen by Releasebot:
Jul 15, 2026
AssemblyAI

Compare Model Responses in the LLM Gateway Playground

AssemblyAI adds side-by-side model response comparison in the LLM Gateway playground and a Vercel AI SDK integration guide.

The LLM Gateway playground can now compare model responses side by side, so you can evaluate how different models answer the same prompt in one view. A Vercel AI SDK integration guide was also added to the docs…
Original source
Jul 10, 2026
Date parsed from source:
Jul 10, 2026

First seen by Releasebot:
Jul 15, 2026
AssemblyAI

Deprecated: Async speech_model Parameter

AssemblyAI updates async speech_model handling for current and legacy model names.

Passing a current model name ( universal-2 , universal-3-5-pro ) to the async speech_model field now returns a 400.

Passing a legacy name ( best , nano , universal ) returns 200 but always routes to the default model…
Original source
Jul 10, 2026
Date parsed from source:
Jul 10, 2026

First seen by Releasebot:
Jul 15, 2026
AssemblyAI

LiveKit Agents: Native Context Carryover

AssemblyAI adds agent_context_carryover to its LiveKit plugin for smoother context carryover in streaming speech-to-text.

LiveKit Agents v1.6.5 adds agent_context_carryover as a first-class parameter on the AssemblyAI plugin, so LiveKit-orchestrated agents can feed conversation context to streaming speech-to-text without the manual wiring…
Original source
Jul 10, 2026
Date parsed from source:
Jul 10, 2026

First seen by Releasebot:
Jul 15, 2026
AssemblyAI

Claude Sonnet 5 on the LLM Gateway

AssemblyAI adds Claude Sonnet 5 in LLM Gateway, with structured outputs unsupported at launch.

Claude Sonnet 5 is now available through the LLM Gateway. Structured outputs are not supported on this model at launch; use tool calling or prompt-based formatting instead…
Original source
Jul 10, 2026
Date parsed from source:
Jul 10, 2026

First seen by Releasebot:
Jul 15, 2026
AssemblyAI

Deprecated: gemini-3-flash-preview Retired from the LLM Gateway

AssemblyAI removes the retired Gemini 3 Flash Preview model from LLM Gateway docs and refreshes stale model recommendations.

The retired gemini-3-flash-preview model has been removed from the LLM Gateway docs and model recommendations, and stale model recommendations were bumped to current versions…
Original source
Jul 10, 2026
Date parsed from source:
Jul 10, 2026

First seen by Releasebot:
Jul 15, 2026
AssemblyAI

GPT-5.6 Luna and Terra on the LLM Gateway

AssemblyAI adds OpenAI GPT-5.6 Luna and Terra to the LLM Gateway with prompt-cache support and dashboard pricing rates.

OpenAI's GPT-5.6 Luna and Terra models are now live on the LLM Gateway, both with prompt-cache support.

Cache read and write rates are listed on the dashboard pricing page…
Original source
Jul 10, 2026
Date parsed from source:
Jul 10, 2026

First seen by Releasebot:
Jul 15, 2026
AssemblyAI

LLM Gateway: Regional and Cache Pricing in the Dashboard

AssemblyAI updates the dashboard pricing table with prompt-cache read and write columns and revised OpenAI and Gemini regional rates.

The dashboard pricing table now shows prompt-cache read and write columns alongside regional rates. Rates for OpenAI and Gemini models were updated: the global tier is set at provider list price, with US and EU…
Original source
Jul 10, 2026
Date parsed from source:
Jul 10, 2026

First seen by Releasebot:
Jul 15, 2026
AssemblyAI

Deprecated: universal-3-pro and universal for New and Inactive Accounts

AssemblyAI restricts Universal-3 Pro requests for new sign-ups and inactive accounts after 7 days.

New sign-ups, and accounts that have not used Universal-3 Pro in the last 7 days, can no longer request universal-3-pro or universal in speech_models. These requests return an error:…
Original source
Jul 9, 2026
Date parsed from source:
Jul 9, 2026

First seen by Releasebot:
Jul 15, 2026
AssemblyAI

Docs: keyterms_prompt Behavior per Model

AssemblyAI updates its accuracy guide to clarify keyterms_prompt behavior across Universal-2 and Universal-3 Pro.

The accuracy guide now states that keyterms_prompt behaves differently on Universal-2 versus Universal-3 Pro, so results match expectations when switching between models…
Original source
Jul 9, 2026
Date parsed from source:
Jul 9, 2026

First seen by Releasebot:
Jul 15, 2026
AssemblyAI

PII Redaction: Granular Location Policies

AssemblyAI adds fine-grained location redaction in PII Redaction with seven granular location values for precise control.

PII Redaction now gives you fine-grained control over location data. redact_pii_policies supports seven granular location values, so you can redact exactly the level of detail you need: location_address ,…
Original source