Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

🔊🔛🔥 ... Groq Inc launched voice generation today. GroqCloud now has realtime transcription, LLMs, and text-to-speech. You can build super-responsive, ultra low-latency voice agents end-to-end entirely on Groq! At Daily, we're big fans of Groq's fast, low-latency inference. Pipecat AI supports all the Groq models, including the new voice... show more

kwindla

14,302 subscribers

31,791 views • 1 year ago •via X (Twitter)

Science & Technology News & Politics

Anya Rossi• Live Now

Private livecam show

11 Comments

kwindla1 year ago

Here's a single-file example showing how to create a realtime voice AI agent using Groq speech-to-text, LLM, and text-to-speech models. ➡️ `pip install 'pipecat-ai[groq,daily,silero]'` ➡️ Configure your API keys ➡️ Run this file: Groq reference docs for the new voice models are here: The voice model is a collaboration between Groq and @play_ht. Their launch blog post is here:

TurboScribe1 year ago

What REAL customers say about TurboScribe's unlimited AI transcription 👇 🇺🇸 "Not only does it transcribe with amazing accuracy, it also filters out a ton of the unnecessary noise associated with pauses in audio. Keep up the great work!" - Kevin

Philipp Tsipman1 year ago

@GroqInc @trydaily Once again, I call on all AI labs and infra vendors to immediately pause for at least 6 months so I, personally, can catch up! 🏳️🏳️🏳️ 😆

PlayAI (formerly PlayHT)1 year ago

@GroqInc @trydaily You should try out Mikail or Mamaw to spice up the game!

James1 year ago

@GroqInc @trydaily which is the best turn detection model? we deeply need a version of whisper that outputs chunks at the 'turn'

kwindla1 year ago

@GroqInc @trydaily Open source and open data turn detection model:

tj1 year ago

@GroqInc @trydaily How many languages supported or only English

kwindla1 year ago

@GroqInc @trydaily English and Arabic.

Lamar1 year ago

@GroqInc @trydaily check this out @turing_hamster

X-Tweets1 year ago

@GroqInc @trydaily

mag1 year ago

@GroqInc @trydaily Which VAD are you using in the demo?

Related Videos

Llama 4 voice agent starter kit with Groq Inc and Pipecat AI ➡️ Groq STT (distil-whisper-large-v3) ➡️ Groq Llama 4 (llama-4-scout-17b-16e-instruct) ➡️ Groq TTS (playai-tts) ➡️ Function calling ➡️ Deploy to Pipecat Cloud for production ➡️ Optionally add a twilio phone number for telephone voice AI

Llama 4 voice agent starter kit with Groq Inc and Pipecat AI ➡️ Groq STT (distil-whisper-large-v3) ➡️ Groq Llama 4 (llama-4-scout-17b-16e-instruct) ➡️ Groq TTS (playai-tts) ➡️ Function calling ➡️ Deploy to Pipecat Cloud for production ➡️ Optionally add a twilio phone number for telephone voice AI

kwindla

58,764 views • 1 year ago

Another example of the multiple TTS parallel pipelines pattern. Here's a voice AI agent that speaks both English and Arabic, using a specific model/voice for each language. These are PlayAI voices. The STT, TTS, and LLM inference is all running on Groq Inc. (The LLM is LLama 4 Maverick.)

Another example of the multiple TTS parallel pipelines pattern. Here's a voice AI agent that speaks both English and Arabic, using a specific model/voice for each language. These are PlayAI voices. The STT, TTS, and LLM inference is all running on Groq Inc. (The LLM is LLama 4 Maverick.)

kwindla

13,401 views • 1 year ago

Very, very fast voice bots. Llama 3.1 running on Groq Inc. 🚀 500ms voice-to-voice response times

Very, very fast voice bots. Llama 3.1 running on Groq Inc. 🚀 500ms voice-to-voice response times

kwindla

386,537 views • 2 years ago

🚨 New model alert! Dialog by vibx — a leading text-to-speech model — now runs on GroqCloud™. That means natural-sounding speech with ultra-low latency, making real-time voice applications smoother and more responsive. Learn more & build fast — links in the comments!

🚨 New model alert! Dialog by vibx — a leading text-to-speech model — now runs on GroqCloud™. That means natural-sounding speech with ultra-low latency, making real-time voice applications smoother and more responsive. Learn more & build fast — links in the comments!

Groq Inc

47,474 views • 1 year ago

NVIDIA just released a new open source transcription model, Nemotron Speech ASR, designed from the ground up for low-latency use cases like voice agents. Here's a voice agent built with this new model. 24ms transcription finalization and total voice-to-voice inference time under 500ms. This agent actually uses *three* NVIDIA open source models: - Nemotron Speech ASR - Nemotron 3 Nano 30GB in a 4-bit quant (released in December) - A preview checkpoint of the upcoming Magpie text-to-speech model These models are all truly open source: weights, training data, training code, and inference code. This is a big deal! Jensen said in the CES keynote yesterday that he expects open source models to catch up to proprietary models this year in a number of categories. NVIDIA is putting their weight behind making this happen. (As Alan Kay said, the best way to predict the future is to invent it.) The code for this agent is open source too, of course. You can deploy it to production with Modal and Pipecat AI cloud, or run locally on an NVIDIA DGX Spark or RTX 5090.

NVIDIA just released a new open source transcription model, Nemotron Speech ASR, designed from the ground up for low-latency use cases like voice agents. Here's a voice agent built with this new model. 24ms transcription finalization and total voice-to-voice inference time under 500ms. This agent actually uses three NVIDIA open source models: - Nemotron Speech ASR - Nemotron 3 Nano 30GB in a 4-bit quant (released in December) - A preview checkpoint of the upcoming Magpie text-to-speech model These models are all truly open source: weights, training data, training code, and inference code. This is a big deal! Jensen said in the CES keynote yesterday that he expects open source models to catch up to proprietary models this year in a number of categories. NVIDIA is putting their weight behind making this happen. (As Alan Kay said, the best way to predict the future is to invent it.) The code for this agent is open source too, of course. You can deploy it to production with Modal and Pipecat AI cloud, or run locally on an NVIDIA DGX Spark or RTX 5090.

kwindla

274,529 views • 6 months ago

How to build the world's fastest voice AI bot: - Self-host speech-to-text, LLM inference, and text-to-speech all together in the same container/cluster. - Route audio over the internet using WebRTC and edge networking. - Configure timings for voice activity detection, phrase endpointing, and other parts of the pipeline to optimize for latency. (There are trade-offs to doing this!) Here's a LLama 3 voice bot that has voice-to-voice response times of ~500ms. We used Deepgram's STT and TTS for this bot, and everything is hosted on cerebriumai's serverless GPU infrastructure.

How to build the world's fastest voice AI bot: - Self-host speech-to-text, LLM inference, and text-to-speech all together in the same container/cluster. - Route audio over the internet using WebRTC and edge networking. - Configure timings for voice activity detection, phrase endpointing, and other parts of the pipeline to optimize for latency. (There are trade-offs to doing this!) Here's a LLama 3 voice bot that has voice-to-voice response times of ~500ms. We used Deepgram's STT and TTS for this bot, and everything is hosted on cerebriumai's serverless GPU infrastructure.

kwindla

282,617 views • 2 years ago

Cerebras inference is very fast. So fast that it changes how we think about configuring our LLMs for voice agent use cases. Kimi K2.6 is a 1T parameter reasoning model that Cerebras serves at 650 - 1,000 tokens per second (end-to-end throughput), with time to first token metrics as low as 150ms (latency). These numbers are two to three times faster than other similarly capable models. The biggest lever we get from this kind of speed is that we can use the model in reasoning mode, and still have excellent "time to first non-thinking token." This solves a big pain point we have in 2026 for voice agent use cases. Almost all recent innovation in post-training has focused on making models good at reasoning ("test time compute"). This is great, but it makes the user-facing model latency much, much slower. Which is a problem for conversational voice agents. We can run Kimi K2.6 with reasoning turned on, and get responses faster than other models produce with reasoning disabled. On my 30-turn voice agent benchmark, Kimi K2.6 with reasoning enabled ties GPT 5.1 and Haiku 4.5 with reasoning disabled, and is still about 200ms seconds faster! On my primary task agent benchmark, Kimi K2.6 is now the #2 model. It ranks just behind Gemini 3.5 Flash in "high" reasoning mode, and tied with GLM 5, Sonnet 4.6, and GPT 5.4 with reasoning set to "low." But Kimi K2.6 completes each turn in the agent loop in under 500ms. The other four models are all at least 3x slower. (Models only qualify for this benchmark if they can complete task turns at a P50 <4s.) A couple of other things that this speed buys us, for production voice agents: - Tool calls happen fast enough that we don't have to work around tool call latency in our pipeline design. - We can prompt the model to output structured data at the beginning of a response, followed by plain text for voice generation. This opens up possibilities like asking the model to do complex classification/generation tasks that influence the rest of the pipeline. For example, the model could create a detailed style prompt for a steerable TTS model, for each individual conversation turn. And, of course, you can use Kimi K2.6 with reasoning turned off. Cerebras calls this "instant" mode. Here's a video of a Cerebras Kimi K2.6 voice agent with voice-to-voice response time, measured at the client, under 500ms. This is the true response latency as perceived by the user, including all network and audio codec overhead, transcription and turn detection, Kimi K2.6 token generation, and voice generation. 500ms is, effectively, instant. So the Cerebras naming for this mode is a propos. :-)

Cerebras inference is very fast. So fast that it changes how we think about configuring our LLMs for voice agent use cases. Kimi K2.6 is a 1T parameter reasoning model that Cerebras serves at 650 - 1,000 tokens per second (end-to-end throughput), with time to first token metrics as low as 150ms (latency). These numbers are two to three times faster than other similarly capable models. The biggest lever we get from this kind of speed is that we can use the model in reasoning mode, and still have excellent "time to first non-thinking token." This solves a big pain point we have in 2026 for voice agent use cases. Almost all recent innovation in post-training has focused on making models good at reasoning ("test time compute"). This is great, but it makes the user-facing model latency much, much slower. Which is a problem for conversational voice agents. We can run Kimi K2.6 with reasoning turned on, and get responses faster than other models produce with reasoning disabled. On my 30-turn voice agent benchmark, Kimi K2.6 with reasoning enabled ties GPT 5.1 and Haiku 4.5 with reasoning disabled, and is still about 200ms seconds faster! On my primary task agent benchmark, Kimi K2.6 is now the #2 model. It ranks just behind Gemini 3.5 Flash in "high" reasoning mode, and tied with GLM 5, Sonnet 4.6, and GPT 5.4 with reasoning set to "low." But Kimi K2.6 completes each turn in the agent loop in under 500ms. The other four models are all at least 3x slower. (Models only qualify for this benchmark if they can complete task turns at a P50 <4s.) A couple of other things that this speed buys us, for production voice agents: - Tool calls happen fast enough that we don't have to work around tool call latency in our pipeline design. - We can prompt the model to output structured data at the beginning of a response, followed by plain text for voice generation. This opens up possibilities like asking the model to do complex classification/generation tasks that influence the rest of the pipeline. For example, the model could create a detailed style prompt for a steerable TTS model, for each individual conversation turn. And, of course, you can use Kimi K2.6 with reasoning turned off. Cerebras calls this "instant" mode. Here's a video of a Cerebras Kimi K2.6 voice agent with voice-to-voice response time, measured at the client, under 500ms. This is the true response latency as perceived by the user, including all network and audio codec overhead, transcription and turn detection, Kimi K2.6 token generation, and voice generation. 500ms is, effectively, instant. So the Cerebras naming for this mode is a propos. :-)

kwindla

40,319 views • 2 months ago

Learn to build conversational AI voice agents in "Building AI Voice Agents for Production", created in collaboration with LiveKit and RealAvatar, and taught by dsa (Co-founder & CEO of LiveKit), Shayne (Developer Advocate, LiveKit), and Nedelina Teneva (Head of AI at RealAvatar, an AI Fund portfolio company). Voice agents combine speech and reasoning capabilities to enable real-time conversations. They're already being used to support customer service, to improve accessibility in healthcare, for entertainment applications, and for talk therapy. In this course, you’ll learn to build voice agents that listen, reason, and respond naturally. You’ll follow the architecture used to create the "AI Andrew" Avatar, a collaborative project between and RealAvatar that responds to users in what sounds like my voice. You’ll build a voice agent from scratch and deploy it to the cloud, enabling support for many simultaneous users. What you’ll learn: - Understand the fundamentals of voice agents, including key components like speech-to-text (STT), text-to-speech (TTS), and LLMs, and how latency is introduced at each layer. - Explore voice agent architectures and the trade-offs between modular pipelines and speech-to-speech APIs. - Explore how platforms like LiveKit mitigate latency issues with optimized networking infrastructure and low-latency communication protocols. - Learn how to connect client devices to voice agents using WebRTC—and why it outperforms HTTP and WebSocket for low-latency audio streaming. - Incorporate voice activity detection (VAD), end-of-turn detection, and context management to detect turns, handle interruptions, and manage conversational flow. - Understand the trade-offs between latency, quality, and cost in an example in which you build a voice agent and change its voice. - Equip your agent with metrics to measure latency at each stage of the voice pipeline and learn the key levers you can pull to make your agent faster and more responsive. The voice agents built in this course also incorporate voice technology from , a supporting contributor to the project. By the end of this course, you'll have learned the components of an AI voice agent pipeline, combined them into a system with low-latency communication, and deployed them on cloud infrastructure so it scales to many users. I’m looking forward to seeing what voice agents you build from this course! Please sign up here:

Andrew Ng

87,484 views • 1 year ago

The voice-to-voice AI Pareto frontier. (You'll never believe this one weird trick ...) If you're building conversational voice AI apps, you care a lot about: ➟ Latency ➟ Cost ➟ LLM response quality (predictable behavior, coverage of the full surface area of your needs, reliable function calling, "reasoning") ➟ Voice quality (correct pronunciations, consistent tone, appropriate affect, steerability, "human-ness") AI model performance is a jagged frontier and has been moving fast for all of 2024. This is especially true for conversational voice, because you're usually using several models in combination. LLMs with native audio capabilities are the newest evolution pushing the performance frontier. Here's a multi-lingual voice conversation using Gemini Flash 1.5's native audio input. But there's a problem ...

The voice-to-voice AI Pareto frontier. (You'll never believe this one weird trick ...) If you're building conversational voice AI apps, you care a lot about: ➟ Latency ➟ Cost ➟ LLM response quality (predictable behavior, coverage of the full surface area of your needs, reliable function calling, "reasoning") ➟ Voice quality (correct pronunciations, consistent tone, appropriate affect, steerability, "human-ness") AI model performance is a jagged frontier and has been moving fast for all of 2024. This is especially true for conversational voice, because you're usually using several models in combination. LLMs with native audio capabilities are the newest evolution pushing the performance frontier. Here's a multi-lingual voice conversation using Gemini Flash 1.5's native audio input. But there's a problem ...

kwindla

23,592 views • 1 year ago

OpenAI shipped a new speech-to-speech model today: gpt-realtime-2 This is the first speech-to-speech model good enough to use in my voice agents that do "real work." Or real play, for that matter. Here's gpt-realtime-2 as the brain of the ship AI in Gradient Bang. The voice-to-voice response and tool calling times here are unedited, so you can see exactly what the interaction with the model is like in an agent with a very complex system instruction and frequent tool calls. (I did clip out the subagent task execution segments, after gpt-realtime-2 starts a subagent via a tool call. Subagents in this config used gpt-5.2 "medium" effort.)

OpenAI shipped a new speech-to-speech model today: gpt-realtime-2 This is the first speech-to-speech model good enough to use in my voice agents that do "real work." Or real play, for that matter. Here's gpt-realtime-2 as the brain of the ship AI in Gradient Bang. The voice-to-voice response and tool calling times here are unedited, so you can see exactly what the interaction with the model is like in an agent with a very complex system instruction and frequent tool calls. (I did clip out the subagent task execution segments, after gpt-realtime-2 starts a subagent via a tool call. Subagents in this config used gpt-5.2 "medium" effort.)

kwindla

54,912 views • 2 months ago

New course: Add voice to your AI agents and applications, built with Vocal Bridge (disclosure: an AI Fund portfolio company) and taught by its CEO Ashwyn Sharma. Voice applications historically required making a hard tradeoff: using fast voice-to-voice models that sacrifice reliability, or accurate speech-to-text pipelines that add latency. This course teaches you how to build voice agents that are both reliable and fast. You'll build three types of voice-enabled applications: a voice-interactive game where voice commands and mouse clicks work together over a single channel, an agent that gains a voice in about 10 lines of code without touching its prompts or tools, and an agent that places outbound phone calls using a make_phone_call function. Skills you'll gain: - Add a voice layer to an existing agent without rewriting your prompts, RAG pipeline, or tools - Give an agent the ability to place outbound calls and stream transcripts back live - Set up voice evaluation to score calls, catch regressions, and improve quality before deployment Join and add voice to your agents without overhauling your architecture:

New course: Add voice to your AI agents and applications, built with Vocal Bridge (disclosure: an AI Fund portfolio company) and taught by its CEO Ashwyn Sharma. Voice applications historically required making a hard tradeoff: using fast voice-to-voice models that sacrifice reliability, or accurate speech-to-text pipelines that add latency. This course teaches you how to build voice agents that are both reliable and fast. You'll build three types of voice-enabled applications: a voice-interactive game where voice commands and mouse clicks work together over a single channel, an agent that gains a voice in about 10 lines of code without touching its prompts or tools, and an agent that places outbound phone calls using a make_phone_call function. Skills you'll gain: - Add a voice layer to an existing agent without rewriting your prompts, RAG pipeline, or tools - Give an agent the ability to place outbound calls and stream transcripts back live - Set up voice evaluation to score calls, catch regressions, and improve quality before deployment Join and add voice to your agents without overhauling your architecture:

Andrew Ng

87,140 views • 1 month ago

Voice-only programming with the new OpenAI Realtime API ... I spend a lot of time these days pair programming with LLMs. Often I'm talking rather than typing. This "voice dictation" use case has become an important vibe benchmark for me. Being able to create text input just by talking, flexibly, in a context dependent way, with tool calling, is a *hard* problem for today's models. Natural language dictation requires a very high degree of contextual intelligence, instruction following accuracy, and tool calling reliability. Today's new gpt-realtime model is quite good at this hard problem. The original realtime model release last year was impressive. Seeing what a speech-to-speech model could do got a lot of people excited about the possibilities of voice AI. The improvements since that first release are equally impressive. I can use this new model, now, for real world tasks that were past the edge of the "jagged frontier" before. Here's a video showing a couple of fun (and tricky) modes of voice input.

Voice-only programming with the new OpenAI Realtime API ... I spend a lot of time these days pair programming with LLMs. Often I'm talking rather than typing. This "voice dictation" use case has become an important vibe benchmark for me. Being able to create text input just by talking, flexibly, in a context dependent way, with tool calling, is a hard problem for today's models. Natural language dictation requires a very high degree of contextual intelligence, instruction following accuracy, and tool calling reliability. Today's new gpt-realtime model is quite good at this hard problem. The original realtime model release last year was impressive. Seeing what a speech-to-speech model could do got a lot of people excited about the possibilities of voice AI. The improvements since that first release are equally impressive. I can use this new model, now, for real world tasks that were past the edge of the "jagged frontier" before. Here's a video showing a couple of fun (and tricky) modes of voice input.

kwindla

50,874 views • 11 months ago

A voice agent powered by gpt-oss. Running locally on my macBook. Demo recorded in a Waymo with WiFi turned off. I'm still on my space game voice AI kick, obviously. Code link below. For conversational voice AI, you want to set the gpt-oss reasoning behavior to "low". (The default is "medium".) Notes on how to do that and a jinja template you can use are in the repo. The LLM in the demo video is the big, 120B version of gpt-oss. You can use the smaller, 20B model for this, of course. But OpenAI really did a cool thing here designing the 120B model to run in "just" 80GB of VRAM. And the llama.cpp mlx inference is fast: ~250ms TTFT. Running a big model on-device feels like a time warp into the future of AI.

A voice agent powered by gpt-oss. Running locally on my macBook. Demo recorded in a Waymo with WiFi turned off. I'm still on my space game voice AI kick, obviously. Code link below. For conversational voice AI, you want to set the gpt-oss reasoning behavior to "low". (The default is "medium".) Notes on how to do that and a jinja template you can use are in the repo. The LLM in the demo video is the big, 120B version of gpt-oss. You can use the smaller, 20B model for this, of course. But OpenAI really did a cool thing here designing the 120B model to run in "just" 80GB of VRAM. And the llama.cpp mlx inference is fast: ~250ms TTFT. Running a big model on-device feels like a time warp into the future of AI.

kwindla

202,147 views • 11 months ago

This is the most realistic voice cloning I’ve ever heard. It captures voice, emotion, speaking style, and even your rhythm. Powered by EVI 3, Hume’s next-gen speech-to-speech model that: ➔ Streams your speech → replies in real-time ➔ Understands tone and context ➔ Adapts emotion on cue (“whisper fearfully,” “sound excited”) Try it now: 🔗 (live demo, no prompt edits) 🔗 (customize voices, integrate with LLMs) You can even integrate it with top AI models like Groq, Anthropic, and DeepSeek 👇 Talk to the voice I cloned here:

This is the most realistic voice cloning I’ve ever heard. It captures voice, emotion, speaking style, and even your rhythm. Powered by EVI 3, Hume’s next-gen speech-to-speech model that: ➔ Streams your speech → replies in real-time ➔ Understands tone and context ➔ Adapts emotion on cue (“whisper fearfully,” “sound excited”) Try it now: 🔗 (live demo, no prompt edits) 🔗 (customize voices, integrate with LLMs) You can even integrate it with top AI models like Groq, Anthropic, and DeepSeek 👇 Talk to the voice I cloned here:

Ali Sufian

15,063 views • 1 year ago

Voice AI turn taking is a solved problem. The single most common complaint about voice AI, today, is that agents interrupt too often. But the voice agents I build for myself now respond quickly and interrupt me less often than the people I talk to every day. (I actually measured this.) Mark Backman made a Pipecat AI PR two weeks ago that was the last piece of the puzzle for turn taking so good that I no longer ever think about it. The approach combines three layers of processing: 1. Voice activity detection, with a short (200ms) trigger. 2. A native audio turn detection model that's small, fast, and runs on CPU. This model captures audio nuances like inflection and filler sounds that don't get transcribed. 3. A prompt mixin for the conversation LLM that decides turn completion based on conversation context. None of these are new. We've been using VAD for a long time. We trained the first version of the Pipecat Smart Turn native audio model in December 2024. And we've been experimenting with prompt-based large model turn detection (sometimes called "selective refusal") for more than a year. Now, the Smart Turn model and the SOTA LLMs we're using in voice agents have both gotten so good that using them together feels like we've finally "solved" turn detection. Mark also figured out how to elegantly apply a "single-token tagging" technique to this problem. We sometimes use single-token tagging in place of tool calling, when we need a near-zero latency programmatic trigger. Mark's Pipecat mixin defines three single-token characters and prompts the LLM to output exactly one of them at the beginning of every response. - ✓ means the agent should respond normally (immediately) - ○ is a "short incomplete" - the agent should wait 5 seconds - ◐ is a "long incomplete" - the agent should wait 10 seconds The wait times, and the details of the prompt, are configurable, of course. Watch the video to see me talk to an agent that handles all my various pauses and inflections, plus phrases like "let me think," pretty much the way a person would handle them, in terms of response latency. Also, in the second half of the video, I ask the agent to adjust its response pattern because I'm going to tell it a phone number. This kind of "in-context" adjustment of response wait times is really useful. The LLM in the video is GTP-4.1. We've tested the prompt and single-token adherance with GPT-4.1, Gemini 2.5 Flash, Anthropic Claude Sonnet 4.5, and AWS Nova 2 Pro. Note that older models in all these families (and, in general, smaller open weights models) aren't able to reliably output these single-token tags. But the new models we're using these days are pretty amazing.

Voice AI turn taking is a solved problem. The single most common complaint about voice AI, today, is that agents interrupt too often. But the voice agents I build for myself now respond quickly and interrupt me less often than the people I talk to every day. (I actually measured this.) Mark Backman made a Pipecat AI PR two weeks ago that was the last piece of the puzzle for turn taking so good that I no longer ever think about it. The approach combines three layers of processing: 1. Voice activity detection, with a short (200ms) trigger. 2. A native audio turn detection model that's small, fast, and runs on CPU. This model captures audio nuances like inflection and filler sounds that don't get transcribed. 3. A prompt mixin for the conversation LLM that decides turn completion based on conversation context. None of these are new. We've been using VAD for a long time. We trained the first version of the Pipecat Smart Turn native audio model in December 2024. And we've been experimenting with prompt-based large model turn detection (sometimes called "selective refusal") for more than a year. Now, the Smart Turn model and the SOTA LLMs we're using in voice agents have both gotten so good that using them together feels like we've finally "solved" turn detection. Mark also figured out how to elegantly apply a "single-token tagging" technique to this problem. We sometimes use single-token tagging in place of tool calling, when we need a near-zero latency programmatic trigger. Mark's Pipecat mixin defines three single-token characters and prompts the LLM to output exactly one of them at the beginning of every response. - ✓ means the agent should respond normally (immediately) - ○ is a "short incomplete" - the agent should wait 5 seconds - ◐ is a "long incomplete" - the agent should wait 10 seconds The wait times, and the details of the prompt, are configurable, of course. Watch the video to see me talk to an agent that handles all my various pauses and inflections, plus phrases like "let me think," pretty much the way a person would handle them, in terms of response latency. Also, in the second half of the video, I ask the agent to adjust its response pattern because I'm going to tell it a phone number. This kind of "in-context" adjustment of response wait times is really useful. The LLM in the video is GTP-4.1. We've tested the prompt and single-token adherance with GPT-4.1, Gemini 2.5 Flash, Anthropic Claude Sonnet 4.5, and AWS Nova 2 Pro. Note that older models in all these families (and, in general, smaller open weights models) aren't able to reliably output these single-token tags. But the new models we're using these days are pretty amazing.

kwindla

26,918 views • 5 months ago

Local voice AI with a 235 billion parameter LLM. ✅ - smart-turn v2 - MLX Whisper (large-v3-turbo-q4) - Qwen3-235B-A22B-Instruct-2507-3bit-DWQ - Kokoro All models running local on an M4 mac. Max RAM usage ~110GB. Voice-to-voice latency is ~950ms. There are a couple of relatively easy ways to carve another ~100ms off that number. But it's not a bad start!

Local voice AI with a 235 billion parameter LLM. ✅ - smart-turn v2 - MLX Whisper (large-v3-turbo-q4) - Qwen3-235B-A22B-Instruct-2507-3bit-DWQ - Kokoro All models running local on an M4 mac. Max RAM usage ~110GB. Voice-to-voice latency is ~950ms. There are a couple of relatively easy ways to carve another ~100ms off that number. But it's not a bad start!

kwindla

66,639 views • 1 year ago

Further tinkering with my little French tutor app. This version is using the Gemini Multimodal Live API. The speech understanding in Gemini is quite something. In this video you can see Gemini correcting my pronunciation. (Very patiently.) The language tutor use case really highlights the strengths of a next-generation speech model like Gemini. This is 90 lines of Pipecat AI code, and uses WebRTC for super low-latency, super reliable network transport.

Further tinkering with my little French tutor app. This version is using the Gemini Multimodal Live API. The speech understanding in Gemini is quite something. In this video you can see Gemini correcting my pronunciation. (Very patiently.) The language tutor use case really highlights the strengths of a next-generation speech model like Gemini. This is 90 lines of Pipecat AI code, and uses WebRTC for super low-latency, super reliable network transport.

kwindla

13,005 views • 1 year ago

Launch day for Google Imagen 3! We updated Story Bot — one of the classic Pipecat AI voice interaction starter kits — to use Gemini 2.0 Flash and Imagen. The new experience with these two models is … wow! Watch Chad's interactive story about a dragon who feels a little bit out of place, and her magical friend. The consistency of the images that Imagen creates through the whole story is a new capability. We haven’t seen that from any other model we’ve experimented with. Clone the repo to build your own voice+images experience.

Launch day for Google Imagen 3! We updated Story Bot — one of the classic Pipecat AI voice interaction starter kits — to use Gemini 2.0 Flash and Imagen. The new experience with these two models is … wow! Watch Chad's interactive story about a dragon who feels a little bit out of place, and her magical friend. The consistency of the images that Imagen creates through the whole story is a new capability. We haven’t seen that from any other model we’ve experimented with. Clone the repo to build your own voice+images experience.

kwindla

22,673 views • 1 year ago

Positively pumped 🫨to announce that raised a $64M Series A lead by Bucky Moore at Kleiner Perkins! - Our latest all-around audio model, Sonic 2.0 has a 90ms TTFA with improved transcript following for the toughest transcripts 🚄 - Our extreme low latency model, Sonic Turbo has a 40ms TTFA. It's so fast it might as well be in the room with you 👀 Elevate your voice AI shitpost game today 🔊

Positively pumped 🫨to announce that raised a $64M Series A lead by Bucky Moore at Kleiner Perkins! - Our latest all-around audio model, Sonic 2.0 has a 90ms TTFA with improved transcript following for the toughest transcripts 🚄 - Our extreme low latency model, Sonic Turbo has a 40ms TTFA. It's so fast it might as well be in the room with you 👀 Elevate your voice AI shitpost game today 🔊

Timothy Luong (Chongz)

50,341 views • 1 year ago

The new Sonic-3 voice model from launched today. The big additions are increased emotional range, emotion steering, and laughter tags. For example: I wrote some quick demo code, prompting Gemini Flash to know about the emotion tags. You can hear the results in the video, and see the emotion tags in the Pipecat developer console. Code is below.

The new Sonic-3 voice model from launched today. The big additions are increased emotional range, emotion steering, and laughter tags. For example: I wrote some quick demo code, prompting Gemini Flash to know about the emotion tags. You can hear the results in the video, and see the emotion tags in the Pipecat developer console. Code is below.

kwindla

12,350 views • 9 months ago