hotwirednews Get GrokBot

Week 30

20–26 Jul 20262026-W30 · 32 filings · 22 companies · 6 net-new

The Use layer led the week with 10 of 32 filings, from Valence AI, SoundHound AI, OpenAI, Anthropic and 5 more.

Hearspeech-to-text3

StenographEnterprisepartnership2026-07-22net-new

Stenograph and Speechmatics put on-device speech recognition into CATalyst VP

This is a named enterprise deployment of on-device speech recognition in a specialist professional workflow, not a general Speechmatics model release.

Details2 sources · 1 X post

CATalyst VP gains on-device, real-time speech recognition for voice-enabled court reporting.

CATalyst VP on-device speech recognition integration with Speechmatics · vendor Speechmatics

SpeechmaticsModel providerlaunch2026-07-22

Speechmatics adds Speaker Focus to keep voice agents on the target speaker

Speaker Focus adds target-speaker control to voice-agent input; the follow-up shows the same feature in support calls, vehicles, and meetings.

Also Orchestrate

Details2 sources · 2 X posts

Speaker Focus locks a voice agent onto the person driving the conversation while treating other nearby voices as background or ignored.

Speaker Focus · vendor LiveKit

It keeps the agent on the person driving the conversation, while other voices stay in the background or are ignored.

https://x.com/Speechmatics/status/2079887552484356363
https://x.com/Speechmatics/status/2082382995462189161
DeepgramModel providermodel release2026-07-21

Deepgram brings Nova-3 on-device speech recognition to Snapdragon X PCs

This extends Nova-3 from hosted speech recognition into local PC inference; the X post is the dated Deepgram announcement and Yahoo carries the related release.

Details2 sources · 1 X post

Nova-3 runs on Qualcomm Hexagon NPU hardware in Snapdragon X Series PCs for real-time speech recognition without a cloud round trip.

Nova-3 on-device for Snapdragon X Series PCs

Voice AI, minus the cloud round trip.

https://x.com/DeepgramAI/status/2079598308834652340

Speaktext-to-speech5

LiveKitPlatformlaunch2026-07-25

LiveKit Agents 1.6.7 adds Speechify streaming TTS with word timestamps

Speechify streaming TTS addition.

Details1 source

livekit-agents 1.6.7 adds Speechify streaming TTS with word-level timestamps and Fish Audio loudness/generation-tuning options.

livekit-agents 1.6.7

MicrosoftModel providermodel release2026-07-23

MAI-Voice-2-Flash: 2× faster TTS at $15 / 1M characters

In-house TTS at $15/1M characters for Dynamics Contact Center and Azure Voice Live. Vendor also claims up to 89% GPU-cost reduction in Contact Center.

Details1 source

Microsoft put MAI-Voice-2-Flash in public preview: 2× faster and 32% cheaper than MAI-Voice-2, at $15 per 1M characters, for Dynamics 365 Contact Center and Azure Voice Live.

MAI-Voice-2-Flash

SpeechifyModel providerlaunch2026-07-22

Simba 3.0 adds five European-language options

Two weeks after Simba 3.2 shipped English-only, Speechify took Simba 3.0 into six languages and began retiring Simba 1.6, which gives the multilingual track its own model line.

Details1 source

Simba 3.0 adds German, Spanish for Spain and Mexico, French, Italian, and Brazilian Portuguese alongside English.

Simba 3.0

Daily / PipecatPlatformlaunch2026-07-21

Pipecat 1.6.0 adds Deepgram Flux TTS

Builders can pair Flux STT and Flux TTS inside one Pipecat cascade.

Details3 sources

pipecat-ai 1.6.0 adds DeepgramFluxTTSService for Deepgram's Flux TTS early access over WebSocket.

pipecat-ai 1.6.0

ByteDance SeedModel providermodel release2026-07-20net-new

ByteDance Seed releases Seed Audio 1.0 for unified scene audio generation

Seed Audio 1.0 makes speech generation part of a unified film and content audio stack rather than a standalone TTS call.

Also Show

Details1 source

Seed Audio 1.0 jointly models voice, sound effects, ambience, and other audio elements for full-scene audio generation.

Seed Audio 1.0

Showavatars, video2

LiveKitPlatformlaunch2026-07-25

LiveKit Agents 1.6.7 adds Spatius avatar and inference.AvatarSession

Avatar gateway provisioning lands with Spatius.

Details1 source

livekit-agents 1.6.7 adds a Spatius avatar plugin and inference.AvatarSession for avatar provisioning via the gateway.

livekit-agents 1.6.7

TavusPlatformlaunch2026-07-20

Tavus gives PALs a Presentation Mode for interactive decks

A conversational avatar can move from answering questions to presenting structured source material while staying in the same live interaction; no independent presentation-quality measure is published.

Also OrchestrateSpeak

Details2 sources · 1 X post

Tavus launched Presentation Mode, letting a PAL present a deck or source material, answer questions, and pull the relevant slide on demand.

Presentation Mode

Your PAL can now present a deck, take questions, and pull the right slide on demand.

https://x.com/tavus/status/2079258477151834420

ThinkLLMs, speech-to-speech2

Daily / PipecatPlatformAPI2026-07-21

Pipecat 1.6.0 adds OpenAI Responses reasoning and Flows NO_RESPONSE

Silent tool outcomes and Responses reasoning controls land for agentic voice flows.

Details3 sources

pipecat-ai 1.6.0 adds reasoning support on OpenAIResponsesLLMService and NO_RESPONSE returns from Pipecat Flows consolidated functions.

pipecat-ai 1.6.0

OpenAIModel providereol2026-07-20

OpenAI sets 20 January 2027 removal for legacy realtime, audio, and transcription models

Speech-to-speech agents still pinned to gpt-realtime or the GPT-4o realtime models have a fixed migration date, and the replacements are newer snapshots in the same families rather than a new API. Framework services that default to older model names need to move with them.

Also Hear

Details1 source

OpenAI notified API developers that the gpt-realtime, gpt-audio, and older GPT-4o realtime and audio families, plus the first gpt-4o-mini-transcribe snapshot, leave the API on 20 January 2027.

Legacy realtime, audio, and transcription snapshots

Orchestrateframeworks, evals7

Daily / PipecatPlatformAPI2026-07-26

Pipecat Context Hub 0.3.0 mounts into the Pipecat CLI and adds a one-step install command

The pipecat mcp name lasted a week: 0.4.0 renamed the bridge (2026-08-02-pipecat-context-hub-0-4-0, W31).

Details2 sources

Pipecat Context Hub 0.3.0 mounts into the Pipecat CLI as pipecat mcp and adds an install command that registers the MCP server with a coding agent and builds the index.

Pipecat Context Hub 0.3.0

LiveKitPlatformlaunch2026-07-25

LiveKit Agents 1.6.7 adds adaptive interruption for realtime models

Realtime interruption parity with pipeline agents. Class-2: folded quiet patch(es) 1.6.9 into this row.

Details2 sources

livekit-agents 1.6.7 supports adaptive interruption for realtime models.

livekit-agents 1.6.7

telliPlatformfunding2026-07-23net-new

telli raises $15M and launches Charlie for building and improving voice agents

The funding and product launch landed together, but Charlie is one product within telli rather than a separately filed company; later telli launch-week rows cover its browser and simulation surfaces.

Also Use

Details2 sources · 1 X post

telli raised a $15 million seed led by Redalpine and launched Charlie, an AI coworker for customer-facing operations.

telli Charlie and $15M seed

We just raised $15M to build AI for customer-facing operations.

https://x.com/tellicom/status/2080232570642440414
VapiPlatformlaunch2026-07-21

Vapi Model Intelligence adds presets and production model metrics

Turns Vapi's model-agnostic catalog into guided defaults, so builders pick a goal (latency, cost, intelligence) instead of hand-picking every STT, LLM, and TTS pair.

Details1 source

Vapi shipped Model Intelligence: curated Model Presets for Balanced, High Intelligence, Ultra Fast, and Cost Saver stacks, plus dashboard cost, latency, and quality metrics drawn from live Vapi traffic.

Vapi Model Intelligence

Daily / PipecatPlatformAPI2026-07-21

Pipecat 1.6.0 adds absent expectations to the eval harness

Negative assertions catch bots that speak when they should stay silent.

Details3 sources

pipecat-ai 1.6.0 lets eval scenario expectations set absent: true so a turn passes only when no matching event arrives within the window.

pipecat-ai 1.6.0

LangChainPlatformlaunch2026-07-21net-new

LangChain adds LangSmith tracing for voice agents across leading realtime stacks

Voice-agent observability now covers the audio and timing boundaries that ordinary text-agent traces miss.

Details2 sources · 1 X post

LangSmith now traces voice agents built with Pipecat, LiveKit, OpenAI Realtime, and Gemini Live, including audio, speech latency, interruptions, and tool calls.

LangSmith voice-agent tracing

Bland AIPlatformlaunch2026-07-21

Bland adds five-level interruptibility and a resumption-speed control for its phone agents

Turn-taking becomes a per-node builder setting instead of one global threshold, the same problem Bland's Fluent endpointing (2026-04-09-bland-fluent-stt) tackles inside the STT; the Talkdesk and NICE CXone BYOC trunks in this release let contact centres keep their CCaaS carrier while running Bland agents.

Also Connect

Details2 sources

Builders set how readily a Bland agent yields when a caller talks over it and how long it waits before replying, per node, per inbound number, or per call through the API.

Agent Speech Timing (Interruptibility and Resumption Speed)

Connecttelephony, channels3

LiveKitPlatformlaunch2026-07-25

LiveKit Agents 1.6.7 adds Krisp voice isolation telephony mode

Telephony noise isolation mode.

Details1 source

livekit-agents 1.6.7 adds Krisp voice isolation telephony mode on the noise-cancellation plugin.

livekit-agents 1.6.7

Daily / PipecatPlatformlaunch2026-07-21

Pipecat 1.6.0 adds Media over QUIC (MoQ) transport

MoQ is a new transport class beside WebRTC and WebSockets for fan-out media.

Details5 sources

pipecat-ai 1.6.0 adds MOQTransport for bidirectional low-latency Media over QUIC between bots and clients.

pipecat-ai 1.6.0

LiveKitPlatformlaunch2026-07-20

LiveKit Unity SDK 2.0.0 leaves developer preview with voice-agent characters for games

AstroBeam's Stellar Cafe is a shipped Meta Quest title on this SDK, a games customer for voice-agent characters rather than a sample. In the same week LiveKit Agents 1.6.7 added avatar sessions (2026-07-25-livekit-agents-1-6-7-show); the blog does not tie the two releases together.

Details2 sources · 1 X post

The Unity client reaches feature parity with LiveKit's other SDKs, so Unity games and VR apps can carry voice chat and video, drive robots over data tracks, and connect players to LiveKit voice agents as non-player characters.

LiveKit Unity SDK 2.0.0

LiveKit Unity SDK 2.0.0 is out of Developer Preview. Bring real-time voice, video, and AI into your Unity game. Build voice chat, live streaming, robot teleoperation, and voice AI agents as NPCs that listen, respond, and react. Give it a try today and let us know what you think. https://t.co/iuLZvcf82k

https://x.com/livekit/status/2079633042868511016

Useagents in market10

Valence AIPlatformfunding2026-07-23net-new

Valence AI raises $5M to add emotional intelligence to voice-AI trust systems

The raise positions emotion inference as a trust and identity layer around voice interactions, rather than as a conversational model launch.

Details1 source

Valence AI raised $5 million backed by SRI to integrate emotional intelligence into identity and fraud-prevention systems.

$5M funding for emotional intelligence in the trust stack

SoundHound AIPlatformpartnership2026-07-23

SoundHound Smart Ordering integrates with Deliverect for end-to-end voice ordering

Turns restaurant voice AI from a standalone answering SKU into kitchen-connected fulfilment via an existing digital-ordering backbone.

Details2 sources

SoundHound and Deliverect connected Smart Ordering voice agents to Deliverect menus and POS routing so drive-thru, phone, kiosk, and in-car orders reach the kitchen without re-keying.

Smart Ordering × Deliverect

OpenAIPlatformlaunch2026-07-23

OpenAI brings ChatGPT Voice to the desktop app with Appshots screen context

Moves GPT-Live from phone chat into a desktop operator loop — Work/Codex steering plus macOS Appshots window capture — two weeks after the 8 July mobile GPT-Live default swap.

Also SeeThink

Details5 sources · 1 X post

OpenAI rolled out ChatGPT Voice in the macOS and Windows desktop app, powered by GPT-Live, so users can speak to steer ChatGPT Work and Codex agents, with macOS Appshots supplying the frontmost window as screen context.

ChatGPT Voice on desktop (Appshots)

ChatGPT Voice is now in the desktop app. Control your computer and direct multiple agents running in ChatGPT Work or Codex, using just your voice.

https://x.com/OpenAI/status/2080378182469857576
AnthropicModel providerlaunch2026-07-23net-new

Claude Voice Mode expands to Opus and Sonnet, still turn-based

Opus and Sonnet on voice without barge-in. Date from @claudeai coverage; tweet URL not independently verified.

Details3 sources

Anthropic expanded the Claude voice mode beta to Opus 5 and Sonnet 5 (plus Haiku 4.5) on mobile, desktop, and web. It remains turn-based, not full-duplex.

Claude Voice Mode

ZoomPlatformlaunch2026-07-22

Zoom launches Voice Translator for real-time spoken translation in meetings

Spoken translation becomes a native meeting capability in a mainstream conferencing product, reducing the need to route multilingual calls through a separate translation service; the launch gives no independent quality benchmark.

Also HearSpeak

Details2 sources · 1 X post

Zoom launched Voice Translator, a meeting feature that translates spoken conversation in real time with five languages at launch.

Zoom Voice Translator

Introducing Zoom Voice Translator.

https://x.com/Zoom/status/2080295465426272676
OpenAIPlatformlaunch2026-07-22

OpenAI launches Presence for governed enterprise voice and chat agents

Moves OpenAI from API voice models into a managed contact-centre-style SKU, competing with Sierra and other agent platforms on governed production deployments rather than only on gpt-realtime and GPT-Live.

Also ThinkOrchestrate

Details3 sources

OpenAI released Presence, a managed enterprise product for deploying governed AI agents on voice and chat that answer questions, take approved actions in company systems, and escalate to people under customer-set policies.

OpenAI Presence

DecagonPlatformlaunch2026-07-22

Decagon makes its AI customer-service platform available on AWS Marketplace

Marketplace availability gives enterprise buyers a cloud-procurement path for Decagon's customer-service agents and resembles other dated marketplace distribution launches in this corpus.

Also Orchestrate

Details2 sources · 1 X post

Decagon listed its AI customer-service platform on AWS Marketplace so AWS customers can procure and deploy it through their existing cloud relationship.

Decagon on AWS Marketplace · delivered by AWS

Decagon is now available on AWS Marketplace.

https://x.com/DecagonAI/status/2079971784829583679
CrestaEnablerpartnership2026-07-22

Cresta and Atento partner to deploy AI customer-service solutions in Latin America

A large BPO channel can put Cresta's voice and customer-service automation into regional operations, making the relationship an operator/enabler signal rather than a generic product teaser.

Also Orchestrate

Details2 sources · 1 X post

Cresta and Atento formed a strategic partnership to bring Cresta's AI customer-service platform to Atento's operations and customers across Latin America.

Cresta and Atento strategic partnership · vendor Cresta · delivered by Atento

Cresta and Atento are partnering to transform customer service in Latin America.

https://x.com/cresta/status/2079963891044421679
TeslaEnterpriselaunch2026-07-21

Tesla's Summer Update lets Grok place phone calls and run car controls by voice

Grok goes from answering questions to acting on the car, two months before Connectors (2026-09-22-tesla-grok-connectors) extend it to email, calendar, and food orders; Tesla publishes no command-success or adoption figures.

Also Think

Details4 sources · 3 X posts

Tesla's Summer Release 2026 (software 2026.26) lets Grok make phone calls, search and play music, adjust climate, open the glovebox, and answer questions about the car.

Grok in Tesla: phone calls and vehicle controls (Summer Release 2026.26) · vendor xAI

https://x.com/Tesla/status/2079584176182087827

Q2 Shareholder Update (excerpt): Automotive and Other Software – Rolled out Summer Release: Self-Driving stats are now available in the mobile app. Grok can make phone calls, search & play music, and adjust climate controls, among other things. Automatic Navigation expands beyond home & work to support any destination based on personal habits and schedule.

https://x.com/Tesla/status/2080021924072366329

Grok speaks many languages & is now deeper integrated with your Tesla

https://x.com/Tesla/status/2081828774078005751
Otter.aiPlatformlaunch2026-07-21

Otter.ai launches Live Assist for real-time call coaching

Otter cited over 35 million users and more than 1 billion meetings transcribed when it launched, but the one to three deals per rep per year figure is a vendor claim.

Also Hear

Details1 source

The coaching agent joins calls on Zoom, Google Meet, Microsoft calls, or the Otter desktop app and shows private guidance as the conversation unfolds.

Live Assist