Run Audio8 ASR Infinite locally and check a first transcript
Download the pinned Audio8 checkpoint, start its local adapted-vLLM backend, stream a permitted recording and verify the saved transcript against the audio.
Follow the story
AI audio
Explore BIG CHANGE reporting on AI audio: what changed, how it works and why it matters.
Latest reporting
23 stories
Mistral launches Voxtral 3 speech models and voice-agent SDK
Mistral's Voxtral 3 links separate speech models, reasoning and tools in a voice-agent SDK. Its docs show a browser build path, prices and phone limits.
Spotify says it kept 50 million spam and “slop” tracks out of recommendations
Spotify says it kept more than 50 million tracks it calls spam or slop out of recommendations in 2026. Its separate 75 million catalog-removal figure was already reported last year.
How to add an ElevenLabs voice FAQ widget to one website page
Set up a dedicated public ElevenLabs FAQ agent, restrict its allowed hostname, embed the voice widget on one page, and check a live conversation, privacy settings and usage.
Build a local Realtime voice agent with read-only tool calls
Build a live browser voice conversation with OpenAI Realtime and two synthetic read-only tools. Complete server, browser and controller code shows call ID matching, safe continuation, interruption and microphone cleanup.
Build a local browser voice conversation with OpenAI Realtime
Build a local browser microphone conversation with OpenAI Realtime over WebRTC. The full server and page code keep the API key on the server and show how to check playback, stop the microphone and handle failures.
What two-channel audio adds to Xiaomi's MiDashengLM-Spatial
Xiaomi researchers released an audio-language model for spatial questions. Its preprint reports stronger directional benchmark scores with two-channel input, while complex scenes remain difficult.
How to use Myna local dictation on Ubuntu
Canonical's Myna preview can dictate into a focused app on supported Ubuntu GNOME Wayland desktops. Follow the documented PPA setup and check its limits.
Grok Voice 3.0 gains on support test, slows first audio
The hosted Voice Agent API now uses Grok Voice 3.0. An independent simulated support score rises, while first-audio time increases and the listed voice price has a separate text charge.
Square launches voice ordering kiosk built around restaurant menu data
Square's new standalone kiosk lets diners order by voice or touch. Its answers depend on restaurant menu data, while setup requires a Plus or Premium plan and live connection.
YouTube Music starts labeling tracks made largely with generative AI
YouTube Music is rolling out a player label for tracks with significant generative AI involvement. The disclosure does not reveal every step of a track's creation.
Deezer plans a search toggle for fully AI-generated music
Deezer says listeners will be able to hide fully AI-generated tracks, albums and related artists from search, then restore them by switching the setting off. The global rollout is planned before the end of 2026.
Michael Smith sentenced to 18 months for bot streams of AI-generated music
Michael Smith received an 18-month sentence after admitting to a wire-fraud conspiracy built on fake music streams. AI-generated songs supplied catalogue volume, while automated accounts produced the plays that generated royalties.
Find the spoken moment before editing a video
Arcmira's n8n workflow returns up to three timestamped transcript passages. Editors still need to check the source, speaker, cut and reuse rights.
Runway's Eleven v4 speech API: request shape and October pricing
Runway's Eleven v4 speech API accepts scripts up to 2,500 characters. This documentation-based guide explains the request, task result, account setup and October 12 price change.
Microsoft’s MAI-Transcribe-2-Streaming: the live transcription path and its limits
Microsoft's new streaming transcription model returns provisional and final text as audio arrives. Its preview API also leaves turn detection and finalization to the client.
Tavus Griffin: what its one-minute video-call result shows
In Tavus-run one-minute calls, 26 of 54 people believed Griffin-Lite was a person. The study, NVIDIA benchmark and limited preview answer different questions.
Tokyo court says a performer's voice can carry publicity rights, but dismisses TikTok deletion claim
A Tokyo court said unauthorized exploitation of a commercially attractive voice can infringe publicity rights. It dismissed Kenjiro Tsuda's TikTok deletion claims after the posts were removed, without deciding whether the clips infringed.