# speech-to-speech

An open-source, modular low-latency voice-agent pipeline connecting voice activity detection, speech-to-text, an LLM, and text-to-speech via WebSocket and WebRTC.

This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles.

## When Should a Voice Agent Speak? End-of-Turn Detection in Turkish

DevFeed: [When Should a Voice Agent Speak? End-of-Turn Detection in Turkish](<https://devfeed.tech/articles/when-should-a-voice-agent-speak-end-of-turn-detection-in-turkish-145522.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25489>)

Author: Ahmet Tu\\u{g}rul Bayrak, Fatma Nur Korkmaz, Mustafa Serta\\c{c} T\\"urkel, Nur Ya\\u{g}mur Tuncer

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [Conversational AI](<https://devfeed.tech/topics/conversational-ai.md>), [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>), [Whisper](<https://devfeed.tech/topics/whisper.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cl](<https://devfeed.tech/tags/cs-cl.md>), [detection](<https://devfeed.tech/tags/detection.md>), [end-of-turn-detection](<https://devfeed.tech/tags/end-of-turn-detection.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [features](<https://devfeed.tech/tags/features.md>), [fine](<https://devfeed.tech/tags/fine.md>), [language](<https://devfeed.tech/tags/language.md>), [language-model](<https://devfeed.tech/tags/language-model.md>), [latency](<https://devfeed.tech/tags/latency.md>), [logistic-regression](<https://devfeed.tech/tags/logistic-regression.md>), [model](<https://devfeed.tech/tags/model.md>), [recording](<https://devfeed.tech/tags/recording.md>), [recordings](<https://devfeed.tech/tags/recordings.md>), [response](<https://devfeed.tech/tags/response.md>), [simulation](<https://devfeed.tech/tags/simulation.md>), [speaker](<https://devfeed.tech/tags/speaker.md>), [study](<https://devfeed.tech/tags/study.md>), [tests](<https://devfeed.tech/tags/tests.md>), [text](<https://devfeed.tech/tags/text.md>), [voice](<https://devfeed.tech/tags/voice.md>)

### AI overview

In Turkish dyadic video calls, a fine-tuned Turkish TurnGPT model distinguished turn ends from mid-utterance pauses better than a zero-shot Turkish language model and a deployed multilingual detector, and was best calibrated among the three text systems. Current-speaker voice-activity timing was similarly informative; combining timing with the fine-tuned text score produced the strongest timing-based results, including against combinations using either text baseline with logistic regression. In a latency simulation, adding the text score to a silence timeout reduced mean response delay while keeping the interruption rate unchanged. The fine-tuned model also retained its advantage over the zero-shot model on automatic Whisper transcripts. Evaluation used recording-disjoint and speaker-disjoint splits.

### Source excerpt

arXiv:2611.25489v1 Announce Type: new Abstract: A voice agent that waits for a fixed silence before replying either interrupts users who pause mid-utterance or responds late to users who have finished. This study examines the decision at the onset of silence, using turn ends and turn-internal pauses extracted from the Real-TurnTurk corpus of Turkish dyadic video calls. A fine-tuned Turkish TurnGPT model, a zero-shot Turkish language model and a deployed multilingual detector are compared, together with voice-activity timing and facial features, under recording-disjoint and speaker-disjoint evaluation with paired cluster bootstrap tests. On held-out recordings, the fine-tuned model separates turn ends from pauses better than both text baselines and is the best calibrated of the three. Timing features of the current speaker are similarly informative, and their combination with the fine-tuned text score gives the best timing-based results; with logistic regression, this combination also outperforms the same combination with either text baseline. In a latency simulation, adding the text score to a silence timeout shortens the mean response delay at the same interruption rate. On automatic Whisper transcripts, the fine-tuned model retains its advantage over the zero-shot model.

## Type With Your Voice on Linux Using These Open Source Local AI Tools

DevFeed: [Type With Your Voice on Linux Using These Open Source Local AI Tools](<https://devfeed.tech/articles/type-with-your-voice-on-linux-using-these-open-source-local-ai-tools-142318.md>)

Original publisher: [Read original article](<https://feed.itsfoss.com/link/24361/17528980/voice-typing-tools-linux>)

Author: Sourav Rudra

Published: 2026-11-19T03:33:49Z

Content type: article

Language: en

Sources: [It's FOSS](<https://devfeed.tech/sources/it-s-foss.md>)

Topics: [Tool](<https://devfeed.tech/topics/tool.md>), [Linux](<https://devfeed.tech/topics/linux.md>), [Open Source](<https://devfeed.tech/topics/open-source.md>), [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-tools](<https://devfeed.tech/tags/ai-tools.md>), [apps](<https://devfeed.tech/tags/apps.md>), [linux](<https://devfeed.tech/tags/linux.md>), [list](<https://devfeed.tech/tags/list.md>), [local](<https://devfeed.tech/tags/local.md>), [local-ai](<https://devfeed.tech/tags/local-ai.md>), [myna](<https://devfeed.tech/tags/myna.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [tools](<https://devfeed.tech/tags/tools.md>), [voice](<https://devfeed.tech/tags/voice.md>)

### AI overview

The listed Linux dictation apps can send speech-to-text into the focused window, and most offer local recognition. Speech Note combines dictation, translation and read-aloud, with X11 support out of the box; Wayland requires the ydotool daemon for typing into other apps. Ubuntu-focused Myna runs locally through separately installed model snaps and is still under heavy development. TuxTalk offers push-to-talk and broad desktop support, while Vocalinux provides voice editing commands and several selectable recognition engines. WhisPaste adds searchable history, vocabulary correction and dictation templates; it also offers optional OpenAI and Deepgram cloud recognition. Hyprwhspr emphasizes scripting, recording modes and hardware-aware local inference, with cloud providers available. Hex keeps local dictation history and custom vocabulary, and also includes screen recording with voice annotation. Installation options and desktop compatibility differ by app; Myna lists specific Ubuntu versions and desktop environments, while TuxTalk notes that users outside Arch-, Debian- or Fedora-based distributions can use its AppImage.

### Source excerpt

The picks range from Ubuntu's Myna to community-built apps that run on a wide variety of distros and desktops.

## How Stretto built an AI voice agent with Amazon Nova 2 Sonic

DevFeed: [How Stretto built an AI voice agent with Amazon Nova 2 Sonic](<https://devfeed.tech/articles/how-stretto-built-an-ai-voice-agent-with-amazon-nova-2-sonic-141098.md>)

Original publisher: [Read original article](<https://aws.amazon.com/blogs/machine-learning/how-stretto-built-an-ai-voice-agent-with-amazon-nova-2-sonic/>)

Author: Mariah Miller

Published: 2026-11-18T18:58:42Z

Content type: article

Language: en

Sources: [Artificial Intelligence](<https://devfeed.tech/sources/artificial-intelligence.md>)

Topics: [Conversational AI](<https://devfeed.tech/topics/conversational-ai.md>), [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>), [Amazon Web Services](<https://devfeed.tech/topics/aws.md>), [Amazon Aurora](<https://devfeed.tech/topics/amazon-aurora.md>)

Tags: [advanced-300](<https://devfeed.tech/tags/advanced-300.md>), [agent](<https://devfeed.tech/tags/agent.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-voice](<https://devfeed.tech/tags/ai-voice.md>), [ai-voice-agent](<https://devfeed.tech/tags/ai-voice-agent.md>), [amazon-aurora](<https://devfeed.tech/tags/amazon-aurora.md>), [amazon-bedrock](<https://devfeed.tech/tags/amazon-bedrock.md>), [amazon-cloudwatch](<https://devfeed.tech/tags/amazon-cloudwatch.md>), [amazon-connect](<https://devfeed.tech/tags/amazon-connect.md>), [amazon-elastic-container-service](<https://devfeed.tech/tags/amazon-elastic-container-service.md>), [amazon-nova](<https://devfeed.tech/tags/amazon-nova.md>), [amazon-sqs](<https://devfeed.tech/tags/amazon-sqs.md>), [anthropic](<https://devfeed.tech/tags/anthropic.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [aws](<https://devfeed.tech/tags/aws.md>), [bankruptcy](<https://devfeed.tech/tags/bankruptcy.md>), [customer-solutions](<https://devfeed.tech/tags/customer-solutions.md>), [generative-ai](<https://devfeed.tech/tags/generative-ai.md>), [interviews](<https://devfeed.tech/tags/interviews.md>), [lead](<https://devfeed.tech/tags/lead.md>), [phone](<https://devfeed.tech/tags/phone.md>), [production](<https://devfeed.tech/tags/production.md>)

### AI overview

Stretto replaced Bankruptcy.com's short web form with Tara, a phone agent that gathers bankruptcy intake details conversationally, then uses attorney-defined rules to sort leads for attorney review. Amazon Nova 2 Sonic handles speech input and output in one model; Claude 3 Haiku extracts typed fields from transcripts, which are stored in Aurora. Twilio streams calls to a voice bridge on ECS Fargate. Because full interviews can exceed one streaming session's duration, the bridge opens a continuation stream and carries over text history. The team validated voice quality, latency, emotional tone, and interruptions before integration, then expanded from the 13-field form to a conditional intake spanning more than 18 categories. Consent and disclosures precede intake; attorneys retain final decisions. After roughly three months, Stretto reported a 20% relative conversion increase, about 96% field extraction accuracy in testing, approximately eight-minute completed calls, and 95% lower per-minute voice costs than its previous solution. Results were early and not yet broadly rolled out.

### Source excerpt

Stretto replaced its web-form lead capture with Tara, an AI voice agent built on Amazon Nova 2 Sonic that conducts bankruptcy intake interviews by phone and qualifies leads. See how Stretto, the AWS Generative AI Innovation Center, and CI&T designed the voice architecture, validated it for sensitive conversations, and moved it to production.

## How Consumer AI Companions Handle Context Windows, Persistent Memory, and Voice Latency

DevFeed: [How Consumer AI Companions Handle Context Windows, Persistent Memory, and Voice Latency](<https://devfeed.tech/articles/how-consumer-ai-companions-handle-context-windows-persistent-memory-and-voice-latency-139410.md>)

Original publisher: [Read original article](<https://www.sitepoint.com/consumer-ai-companions-context-memory-voice-latency/>)

Author: SitePoint Sponsors

Published: 2026-11-17T07:00:00Z

Content type: article

Language: en

Sources: [SitePoint](<https://devfeed.tech/sources/sitepoint.md>)

Topics: [AI Context Management](<https://devfeed.tech/topics/ai-context-management.md>), [Retrieval Augmented Generation (RAG)](<https://devfeed.tech/topics/retrieval-augmented-generation-rag.md>), [Real-time AI interactions](<https://devfeed.tech/topics/real-time-ai-interactions.md>), [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>)

Tags: [2026](<https://devfeed.tech/tags/2026.md>), [404-media](<https://devfeed.tech/tags/404-media.md>), [4k](<https://devfeed.tech/tags/4k.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-companions](<https://devfeed.tech/tags/ai-companions.md>), [apps](<https://devfeed.tech/tags/apps.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [consumer](<https://devfeed.tech/tags/consumer.md>), [consumer-ai](<https://devfeed.tech/tags/consumer-ai.md>), [context](<https://devfeed.tech/tags/context.md>), [handle](<https://devfeed.tech/tags/handle.md>), [latency](<https://devfeed.tech/tags/latency.md>), [llm](<https://devfeed.tech/tags/llm.md>), [memory](<https://devfeed.tech/tags/memory.md>), [persistent-memory](<https://devfeed.tech/tags/persistent-memory.md>), [stateful](<https://devfeed.tech/tags/stateful.md>), [streaming](<https://devfeed.tech/tags/streaming.md>), [voice](<https://devfeed.tech/tags/voice.md>), [windows](<https://devfeed.tech/tags/windows.md>), [with](<https://devfeed.tech/tags/with.md>)

### AI overview

Consumer AI companions maintain continuity by combining a bounded context window with persistent memory: a running summary, extracted user facts, and retrieval of relevant older chats. These layers help but summaries can drift, and retention varies by plan; the article cites a free tier that deletes chat history after seven days of inactivity. Voice latency falls when transcription, LLM output, and text-to-speech stream in sequence, with completed sentences sent to speech synthesis before the full reply is ready; first audio can approach a second on a good connection. Voice also costs more because it uses three model calls and audio bandwidth, so examples include per-minute charges or voice limited to top tiers. The article's practical guidance is to budget prompt space, make stored facts inspectable, stream each voice stage, and protect memory stores because they contain sensitive data.

### Source excerpt

See how consumer AI companions manage context windows, persistent memory and streaming voice, with practical architecture lessons for stateful LLM apps. Continue reading How Consumer AI Companions Handle Context Windows, Persistent Memory, and Voice Latency on SitePoint.

## Does the Voice Matter? Evaluating Frustration Detection on Deployed Clinical Voice-Agent Calls

DevFeed: [Does the Voice Matter? Evaluating Frustration Detection on Deployed Clinical Voice-Agent Calls](<https://devfeed.tech/articles/does-the-voice-matter-evaluating-frustration-detection-on-deployed-clinical-voice-agent-calls-137644.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.21001>)

Author: Annika Michael, Zachary Ellis, Yuxiang Zhou, Samuel Thio, Yajie Vera He, Spencer Hazel, Jared Joselowitz

Published: 2026-11-17T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [Conversational AI](<https://devfeed.tech/topics/conversational-ai.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [asr](<https://devfeed.tech/topics/asr.md>), [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>)

Tags: [audio](<https://devfeed.tech/tags/audio.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cl](<https://devfeed.tech/tags/cs-cl.md>), [detection](<https://devfeed.tech/tags/detection.md>), [language-models](<https://devfeed.tech/tags/language-models.md>), [speech-to-text](<https://devfeed.tech/tags/speech-to-text.md>), [transcript](<https://devfeed.tech/tags/transcript.md>), [triage](<https://devfeed.tech/tags/triage.md>), [voice](<https://devfeed.tech/tags/voice.md>), [voice-agents](<https://devfeed.tech/tags/voice-agents.md>)

### AI overview

In 30 deployed clinical voice-agent calls, patient turns were often too short for transcripts alone to reliably detect frustration: about 59% were one or two words, and transcript-based detection flagged only 3 or 4 of 39 frustrated one-word turns. Adding audio did not consistently improve results. Two Gemini 2.5 models gained 3.0 percentage points when prompted to attend to vocal delivery, but no model showed a statistically clear audio advantage when given the same prompt for both inputs; some models performed worse with audio under input-specific prompts. A WavLM probe reached 71.6% overall accuracy yet missed every frustrated one-word turn in 24 added calls. The findings indicate that a short transcript cannot rule out frustration, while audio benefits depend on the model and prompt.

### Source excerpt

arXiv:2611.21001v1 Announce Type: cross Abstract: Autonomous voice agents now conduct routine clinical follow-up calls, so some automated triage is unavoidable. A patient becoming frustrated is one thing that triage should catch. These systems typically read only the transcript the speech-to-text stage produces, not the voice, and whether that is enough is largely untested on real clinical speech. We study per-turn frustration detection on real patient calls from a deployed clinical voice agent: 150 turns from six calls labelled by three annotators, with a length-matched acted corpus as a control, and 714 turns from 30 calls. Patient turns are overwhelmingly short: across the deployed calls, roughly 59% are one or two words, against 13% in the full acted corpus, and a corpus accuracy gap remains even after the acted set is length-matched. On one-word turns, transcript-only detection does not exclude chance. Across five audio-capable language models, the transcript flags 3 or 4 of the 39 frustrated one-word turns in the 30 calls, and a classifier given only the word count reaches 57.3%. Whether audio helps depends on the model and on the instruction. Asked to attend to delivery, two Gemini 2.5 models gain 3.0 pp over the transcript across 30 calls (-1.0 to 7.5 pp). Of the turns they confidently clear, 6.9% were in fact frustrated for audio against 13.0% for the transcript, a difference of 6.1 pp (2.2 to 10.4 pp). Across the 30 calls, given one prompt identical for both inputs, no model's audio-text interval excludes zero. Under the input-specific prompts, a newer Gemini model and two smaller open models do worse with audio than with the transcript (-3.5, -8.6 and -10.6 pp). A WavLM probe trained on the clinical calls matches the best prompted model overall (71.6%) but misses every frustrated one-word turn in the 24 added calls. When a patient's turn is short, the transcript alone is not enough to rule out frustration.

## WaveGen: Truly End-to-End Waveform Generation via Internal Spectral Trajectory Learning

DevFeed: [WaveGen: Truly End-to-End Waveform Generation via Internal Spectral Trajectory Learning](<https://devfeed.tech/articles/wavegen-truly-end-to-end-waveform-generation-via-internal-spectral-trajectory-learning-126286.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.12527>)

Author: Sang-Hoon Lee, Ha-Yeong Choi

Published: 2026-11-10T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>), [Neural Network](<https://devfeed.tech/topics/neural-network.md>), [scaling laws](<https://devfeed.tech/topics/scaling-laws.md>), [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>)

Tags: [alignment](<https://devfeed.tech/tags/alignment.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-sd](<https://devfeed.tech/tags/cs-sd.md>), [data](<https://devfeed.tech/tags/data.md>), [diffusion](<https://devfeed.tech/tags/diffusion.md>), [eess-as](<https://devfeed.tech/tags/eess-as.md>), [eess-sp](<https://devfeed.tech/tags/eess-sp.md>), [generation](<https://devfeed.tech/tags/generation.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [performance](<https://devfeed.tech/tags/performance.md>), [scalability](<https://devfeed.tech/tags/scalability.md>), [scaling](<https://devfeed.tech/tags/scaling.md>), [source-code](<https://devfeed.tech/tags/source-code.md>), [speech](<https://devfeed.tech/tags/speech.md>), [training-dataset](<https://devfeed.tech/tags/training-dataset.md>), [tts](<https://devfeed.tech/tags/tts.md>)

### AI overview

WaveGen is a diffusion-based text-to-waveform TTS model that learns an internal spectral trajectory alongside the target waveform in one model. Its Spectral Forcing architecture uses disentangled diffusion heads and a multi-scale spectral v-loss; it also handles text-speech alignment internally. The approach avoids separately trained or externally decoded codecs, mel-spectrogram vocoders, VAE latents, duration predictors, self-supervised text or speech representations, and semantic distillation. The authors scale the model from 0.1B to 0.6B parameters and report state-of-the-art performance on the Emilia training dataset with open-source data and one-stage training. On Seed-en, they report WER 1.47 and SIM 0.75.

### Source excerpt

arXiv:2611.12527v1 Announce Type: cross Abstract: This work proposes WaveGen, a fully end-to-end diffusion-based text-to-waveform framework based on Spectral Forcing, which jointly learns an internal spectral trajectory and the target waveform trajectory within a single model, enabling direct end-to-end text-to-waveform generation. Unlike recent TTS pipelines, WaveGen does not rely on separately trained or externally decoded intermediate representations, such as neural audio codecs, Mel-spectrogram vocoders, or VAE-based latent models. In addition, WaveGen performs internal text-speech alignment within a single model, eliminating the need for external alignment modules such as duration predictors. To preserve a truly end-to-end formulation, our framework further avoids dependence on self-supervised representations, such as BERT-based text embeddings and Wav2Vec 2.0-based semantic representations, as well as semantic distillation methods for training optimization. To achieve this, we carefully design Spectral Forcing, a model architecture with disentangled diffusion heads, and a training objective based on a multi-scale spectral $v$-loss. We further validate the scalability of end-to-end TTS models by scaling the model size from 0.1B to 0.6B parameters. Using a single architecture, open-source data, and one-stage training, without external modules such as duration predictor or REPA, WaveGen demonstrates fully end-to-end waveform generation achieving state-of-the-art performance on the Emilia training dataset. In particular, WaveGen achieves a WER of 1.47 and a SIM of 0.75 on the Seed-en benchmark. Source code and checkpoints are available at https://github.com/sh-lee-prml/WaveGen

## Soniox STT v6 Adds Agent Text Input and Echo Filtering for Voice Agents

DevFeed: [Soniox STT v6 Adds Agent Text Input and Echo Filtering for Voice Agents](<https://devfeed.tech/articles/soniox-stt-v6-adds-agent-text-input-and-echo-filtering-for-voice-agents-115949.md>)

Original publisher: [Read original article](<https://www.marktechpost.com/2026/11/03/soniox-stt-v6-agent-text-input-echo-cancellation/>)

Author: Michal Sutter

Published: 2026-11-03T20:18:22Z

Content type: release

Language: en

Sources: [MarkTechPost](<https://devfeed.tech/sources/marktechpost.md>)

Topics: [asr](<https://devfeed.tech/topics/asr.md>), [Conversational AI](<https://devfeed.tech/topics/conversational-ai.md>), [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agent-context](<https://devfeed.tech/tags/agent-context.md>), [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [ai-webinars](<https://devfeed.tech/tags/ai-webinars.md>), [api](<https://devfeed.tech/tags/api.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [audio-language-model](<https://devfeed.tech/tags/audio-language-model.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [cancellation](<https://devfeed.tech/tags/cancellation.md>), [cloud](<https://devfeed.tech/tags/cloud.md>), [echo](<https://devfeed.tech/tags/echo.md>), [editors-pick](<https://devfeed.tech/tags/editors-pick.md>), [enterprise-ai](<https://devfeed.tech/tags/enterprise-ai.md>), [harness](<https://devfeed.tech/tags/harness.md>), [input](<https://devfeed.tech/tags/input.md>), [language-model](<https://devfeed.tech/tags/language-model.md>), [model](<https://devfeed.tech/tags/model.md>), [new-releases](<https://devfeed.tech/tags/new-releases.md>), [release](<https://devfeed.tech/tags/release.md>), [replies](<https://devfeed.tech/tags/replies.md>), [speech](<https://devfeed.tech/tags/speech.md>), [speech-to-text](<https://devfeed.tech/tags/speech-to-text.md>), [speech-to-text-api](<https://devfeed.tech/tags/speech-to-text-api.md>), [staff](<https://devfeed.tech/tags/staff.md>), [streaming](<https://devfeed.tech/tags/streaming.md>), [stt](<https://devfeed.tech/tags/stt.md>), [tech-news](<https://devfeed.tech/tags/tech-news.md>), [technology](<https://devfeed.tech/tags/technology.md>), [text](<https://devfeed.tech/tags/text.md>), [voice](<https://devfeed.tech/tags/voice.md>), [voice-agents](<https://devfeed.tech/tags/voice-agents.md>), [voice-ai](<https://devfeed.tech/tags/voice-ai.md>)

### AI overview

Soniox STT v6 can use a voice agent's streamed text to help recognize short caller replies and filter leaked agent speech. Agent text input is available only on streaming `stt-rt-v6`: send text over the caller-audio WebSocket, ideally 500 ms to 1 second before playback because messages have no timestamps. The feature complements acoustic echo cancellation, which should remain enabled, and still allows callers to interrupt. The file model `stt-async-v6` does not support agent text input. Soniox reports improvements but has published no v6 accuracy or latency results; the cited Pipecat figures are for v5. The excerpt lists prices of $0.12/hour for streaming and $0.10/hour for async, with agent text tokens billed separately.

### Source excerpt

Soniox has released Soniox STT v6, the latest generation of its speech-to-text API. Soniox STT v6 is a speech-to-text model that reads what a voice agent says while hearing the caller. It uses that text to recognize short replies and to drop the agent's own voice. The release ships as stt-rt-v6 for streaming and stt-async-v6 [...] The post Soniox STT v6 Adds Agent Text Input and Echo Filtering for Voice Agents appeared first on MarkTechPost.

## ReFrame: Silence-Aware Compute Reclamation for Full-Duplex Speech Models

DevFeed: [ReFrame: Silence-Aware Compute Reclamation for Full-Duplex Speech Models](<https://devfeed.tech/articles/reframe-silence-aware-compute-reclamation-for-full-duplex-speech-models-114612.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.01835>)

Author: Zheng Zhou, Guanyi Chen, Jing Wei, Haoyuan Liang, Jian Zhang, Haohuan Fu

Published: 2026-11-03T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>), [Retrieval Augmented Generation (RAG)](<https://devfeed.tech/topics/retrieval-augmented-generation-rag.md>), [Transformers](<https://devfeed.tech/topics/transformers.md>), [tgi](<https://devfeed.tech/topics/tgi.md>)

Tags: [compute](<https://devfeed.tech/tags/compute.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [gpu](<https://devfeed.tech/tags/gpu.md>), [language-model](<https://devfeed.tech/tags/language-model.md>), [model-inference](<https://devfeed.tech/tags/model-inference.md>), [retrieval](<https://devfeed.tech/tags/retrieval.md>), [speech](<https://devfeed.tech/tags/speech.md>), [temporal](<https://devfeed.tech/tags/temporal.md>), [tests](<https://devfeed.tech/tags/tests.md>)

### AI overview

ReFrame reclaims compute in full-duplex speech inference by using dialogue state to schedule work. During mutual silence, Mutual Silence Speculation (MSS) evaluates multiple Temporal continuations and verifies stability before correcting Mimi audio tokens in batches. While the user speaks and the agent listens, Listening Frame Folding (LFF) processes arrived frames in parallel and accepts only a verified prefix. Frame-Budget Reinvestment (FBR) uses released intervals for resumable retrieval and language-model workers, then supplies references through Moshi's conditioning interface. Across 1,225 FDBench conversations on five models, the paper reports 18.80-31.01% lower measured GPU execution time with 96.92-100% non-PAD text agreement. In HaluEval tests, FBR raised correct spoken answers from 86 to 205 under the same auxiliary-service budget.

### Source excerpt

arXiv:2611.01835v1 Announce Type: new Abstract: Full-duplex speech models execute dense model inference for every audio frame throughout a live interaction. This uniform framewise schedule creates sustained GPU load across dialogue states whose computation has strong, reusable temporal structure. ReFrame formulates full-duplex inference as a silence-aware computation problem. Mutual Silence Speculation (MSS) operates when both participants are silent: it evaluates multi-frame Temporal continuations, verifies text and hidden-state stability, and applies batched Mimi audio-token correction. Listening Frame Folding (LFF) operates while the user speaks and the agent listens: it processes already-arrived user frames through parallel Temporal execution and frame-batched audio-token generation while accepting only the verified prefix. Frame-Budget Reinvestment (FBR) assigns the released frame intervals to resumable external retrieval and language-model workers, then injects timely references through Moshi's conditioning interface. We instantiate ReFrame on five full-duplex models spanning Moshi-family, phase-structured, and flattened autoregressive generation. Across the complete 1,225-conversation FDBench, ReFrame reduces measured GPU execution time by 18.80-31.01% and retains 96.92-100% non-PAD text agreement on five different models. On HaluEval tests, FBR increases the number of correct generated spoken answers from 86 to 205 under the same auxiliary-service budget. ReFrame turns recurring dialogue structure into lower inference cost and higher end-to-end service capability.

## TamilEOT: A Dataset and Model for Semantic End-of-Turn Detection in Tamil Telephone Speech

DevFeed: [TamilEOT: A Dataset and Model for Semantic End-of-Turn Detection in Tamil Telephone Speech](<https://devfeed.tech/articles/tamileot-a-dataset-and-model-for-semantic-end-of-turn-detection-in-tamil-telephone-speech-111916.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2609.05631>)

Author: Santhoshkumar V

Published: 2026-11-02T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [Advanced AI workflows](<https://devfeed.tech/topics/advanced-ai-workflows.md>), [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>), [Data Quality](<https://devfeed.tech/topics/data-quality.md>), [Local AI](<https://devfeed.tech/topics/local-ai.md>)

Tags: [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cl](<https://devfeed.tech/tags/cs-cl.md>), [cs-sd](<https://devfeed.tech/tags/cs-sd.md>), [dataset](<https://devfeed.tech/tags/dataset.md>), [detection](<https://devfeed.tech/tags/detection.md>), [end-of-turn-detection](<https://devfeed.tech/tags/end-of-turn-detection.md>), [model](<https://devfeed.tech/tags/model.md>), [semantic](<https://devfeed.tech/tags/semantic.md>), [speech](<https://devfeed.tech/tags/speech.md>), [telephone](<https://devfeed.tech/tags/telephone.md>)

### AI overview

TamilEOT introduces an open dataset and two audio-only semantic end-of-turn detectors for Tamil telephone speech: 18,485 labelled boundaries from 116 conversations and a companion set of 4,774 turns. On 4,168 held-out clips from 30 unseen calls, accuracy improves from 70.30% zero-shot to 83.71% and 86.13% for models of 8.7 MB and 21 MB; both run in under 150 ms on a single-threaded laptop CPU. The authors report that rule-derived negative labels agreed with a blind human review only 44.4%, while an audio-LLM labeller reached 97.5% agreement for US$5.69. Of six training changes tested, only encoder capacity improved results beyond the 0.87-point spread across identical runs. Production VAD and streaming replay reduced accuracy by another 2.60 points, and 7.8% of boundaries did not reach the model. The dataset, weights and code are public.

### Source excerpt

arXiv:2609.05631v2 Announce Type: replace-cross Abstract: A voice agent must decide, at each pause, whether the speaker has finished or is still thinking. Without a model of the language, systems fall back to a fixed silence timeout: too short, and the agent interrupts; too long, and every turn pays the full wait. We release TamilEOT, to our knowledge the first open dataset and detector for semantic end-of-turn detection in a South Indian language: 18,485 labelled turn boundaries from 116 real Tamil telephone conversations, two audio-only detectors built on Smart Turn v3's architecture, and a turn-level companion set of 4,774 whole turns with their pause structure. On a held-out split of 4,168 clips from 30 unseen calls, accuracy rises from 70.30% zero-shot to 83.71% (8.7 MB) and 86.13% (21 MB); ROC-AUC rises from 0.751 to 0.921. Both run in under 150 ms single-threaded on a laptop CPU. We also report what building it cost. Our rule-derived negative-class labels scored 44.4% against a blind human pass, below chance, because the rule answered a different question than the model is asked; an audio-LLM labeller at 97.5% agreement replaced them for US$5.69. Of six training levers, only encoder capacity moved the result: three identical runs spanned 0.87 accuracy points, the floor below which nothing else here is a finding. Replaying labelled boundaries through the production VAD and streaming adapter costs a further 2.60 points, and 7.8% of boundaries never reach the model at all. Data, weights, code and every negative result are public.

## The Night Before the Demo, My AI Drive-Thru Started to Stutter

DevFeed: [The Night Before the Demo, My AI Drive-Thru Started to Stutter](<https://devfeed.tech/articles/the-night-before-the-demo-my-ai-drive-thru-started-to-stutter-100951.md>)

Original publisher: [Read original article](<https://proandroiddev.com/building-an-ai-drive-thru-agent-in-flutter-ec3953087ffd?source=rss----c72404660798---4>)

Author: Akashshahriar

Published: 2026-10-26T23:26:05Z

Content type: article

Language: en

Sources: [ProAndroidDev](<https://devfeed.tech/sources/proandroiddev-medium.md>)

Topics: [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>), [asr](<https://devfeed.tech/topics/asr.md>), [WebSocket](<https://devfeed.tech/topics/websocket.md>), [AI Context Management](<https://devfeed.tech/topics/ai-context-management.md>), [User interface design](<https://devfeed.tech/topics/ui-design.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [also](<https://devfeed.tech/tags/also.md>), [android](<https://devfeed.tech/tags/android.md>), [android-app-development](<https://devfeed.tech/tags/android-app-development.md>), [androiddev](<https://devfeed.tech/tags/androiddev.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [audio](<https://devfeed.tech/tags/audio.md>), [brain](<https://devfeed.tech/tags/brain.md>), [can](<https://devfeed.tech/tags/can.md>), [car](<https://devfeed.tech/tags/car.md>), [card](<https://devfeed.tech/tags/card.md>), [checkout](<https://devfeed.tech/tags/checkout.md>), [code](<https://devfeed.tech/tags/code.md>), [counter](<https://devfeed.tech/tags/counter.md>), [deepgram](<https://devfeed.tech/tags/deepgram.md>), [echo](<https://devfeed.tech/tags/echo.md>), [engine](<https://devfeed.tech/tags/engine.md>), [flutter](<https://devfeed.tech/tags/flutter.md>), [full](<https://devfeed.tech/tags/full.md>), [headset](<https://devfeed.tech/tags/headset.md>), [i](<https://devfeed.tech/tags/i.md>), [inside](<https://devfeed.tech/tags/inside.md>), [job](<https://devfeed.tech/tags/job.md>), [just](<https://devfeed.tech/tags/just.md>), [voice-assistant](<https://devfeed.tech/tags/voice-assistant.md>)

### AI overview

The kiosk's menu and cart stayed reliable with a small OpenAI model calling guarded Dart tools, while the hardest problem was audio on the actual Android tablet. Deepgram streaming transcription used tuned pause thresholds, selective menu hints and interim text as a fallback; code enforced stock, quantity and checkout limits, and kept the live cart authoritative. Full-duplex audio worked in development, but the tablet's hardware echo-cancellation output path made speech stutter, and larger buffers added too much delay. The author shipped half-duplex instead, pausing the mic during speech; writing 20 ms of silence into an empty playback buffer kept the track awake and reduced startup buffering to about 250 ms. Checkout remained in the existing kiosk flow after spoken confirmation. The account recommends testing on the real device early; software echo cancellation is a possible next step, with playback-to-microphone delay alignment still unresolved.

### Source excerpt

Deepgram for ears, a small LLM with tools for the brain, a native audio engine for the mouth. How a kiosk takes a full order by voice, and why the echo canceller lost. You know this one. You pull up to the drive-thru speaker. It crackles. "Welcome, what can I get you?" You order a burger, no onions, a large fries and a Coke. Silence. Then: "Sorry, can you repeat that?" You repeat it, louder, leaning out of the window. Behind you a car honks. At the window they hand you a bag with onions in it and no fries. Nobody in that story did anything wrong. The person on the headset is also bagging orders, watching a timer and taking a second lane. The speaker is cheap, the lane is noisy, and the job is to understand fifty strangers an hour through both. Now the other version. You walk up to a kiosk and just talk to it, the way you would talk to a person. "The classic Smashburger. Actually, make it two. And a medium Coke." It answers in a second. It knows the menu, knows what is sold out, never forgets the "no onions", never gets tired at the end of a shift, and reads the order back before it takes your card. You tap, a receipt prints, you leave. Every restaurant chain wants that second version. Some have already tried it in public and pulled it after videos of the AI adding hundreds of nuggets to one order went viral. It is harder than it looks. I built one. A Flutter kiosk that greets you, takes the whole order by voice, reads it back, sends you to the card terminal and prints the receipt. No screen taps until the card. This is how it works, what nearly broke it the day before it ran live, and why the simple version won. What is inside Three parts, the same three you would use to describe a person at the counter: Ears. Deepgram, streaming speech to text over a WebSocket. Brain. A small OpenAI model (gpt-5.4-mini) with tools for the menu, the cart and checkout. Mouth. A native audio plugin that owns the microphone and the speaker. The code below is rewritten and trimmed. The

## Retell AI and Voice Agent APIs

DevFeed: [Retell AI and Voice Agent APIs](<https://devfeed.tech/articles/retell-ai-and-voice-agent-apis-98753.md>)

Original publisher: [Read original article](<https://www.swytchcode.com/content/retell-ai-api-voice-agents>)

Author: Chaitrali Kakde

Published: 2026-10-24T08:00:00Z

Content type: tutorial

Language: en

Sources: [Swytchcode](<https://devfeed.tech/sources/swytchcode.md>)

Topics: [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>), [human in the loop agent](<https://devfeed.tech/topics/human-in-the-loop-agent.md>), [API Economy](<https://devfeed.tech/topics/api-economy.md>)

Tags: [2](<https://devfeed.tech/tags/2.md>), [3](<https://devfeed.tech/tags/3.md>), [about](<https://devfeed.tech/tags/about.md>), [account](<https://devfeed.tech/tags/account.md>), [accounting](<https://devfeed.tech/tags/accounting.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [apis](<https://devfeed.tech/tags/apis.md>), [call](<https://devfeed.tech/tags/call.md>), [change](<https://devfeed.tech/tags/change.md>), [first](<https://devfeed.tech/tags/first.md>), [minutes](<https://devfeed.tech/tags/minutes.md>), [retell](<https://devfeed.tech/tags/retell.md>), [settings](<https://devfeed.tech/tags/settings.md>), [stalled](<https://devfeed.tech/tags/stalled.md>), [up](<https://devfeed.tech/tags/up.md>), [use-cases](<https://devfeed.tech/tags/use-cases.md>), [voice](<https://devfeed.tech/tags/voice.md>), [webhooks](<https://devfeed.tech/tags/webhooks.md>), [with](<https://devfeed.tech/tags/with.md>)

### AI overview

Retell's default silence timeout is ten minutes and its maximum call duration is one hour; the article recommends setting both to limits suited to the expected conversation. It advises grouping failed calls by `disconnection_reason` and treating `registered` as call setup, not proof that someone answered. Webhooks time out after ten seconds by default, so acknowledge promptly, queue processing, and deduplicate by call ID. For outbound calls, `/v2/create-phone-call` supports an `idempotency_key` that prevents duplicate calls on retries for up to an hour. The article recommends approval gates for calls to real numbers, using do-not-call handling, and keeping call creation out of exploratory MCP sessions while allowing read access. It distinguishes Retell's live-call orchestration from ElevenLabs' speech synthesis.

### Source excerpt

The default is ten minutes of silence before a stalled agent hangs up, with a real person on the line. Retell's webhooks, call states, and the settings to change first.

## How to Keep ADK Voice Agents Responsive During Long-Running Tool Calls

DevFeed: [How to Keep ADK Voice Agents Responsive During Long-Running Tool Calls](<https://devfeed.tech/articles/step-by-step-guide-build-a-voice-agent-that-never-goes-silent-using-adk-90417.md>)

Original publisher: [Read original article](<https://developers.googleblog.com/step-by-step-guide-build-a-voice-agent-that-never-goes-silent-using-adk/>)

Author: Stephen Allen; Kaz Sato

Published: 2026-10-20T02:35:00.194306Z

Content type: tutorial

Language: en

Sources: [Google Developers Blog](<https://devfeed.tech/sources/google-developers-blog.md>)

Topics: [google-adk](<https://devfeed.tech/topics/google-adk.md>), [Conversational AI](<https://devfeed.tech/topics/conversational-ai.md>), [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>)

Tags: [adk](<https://devfeed.tech/tags/adk.md>), [async](<https://devfeed.tech/tags/async.md>), [asyncio](<https://devfeed.tech/tags/asyncio.md>), [gemini](<https://devfeed.tech/tags/gemini.md>), [gemini-3-8-live](<https://devfeed.tech/tags/gemini-3-8-live.md>), [guide](<https://devfeed.tech/tags/guide.md>), [python](<https://devfeed.tech/tags/python.md>), [scheduling](<https://devfeed.tech/tags/scheduling.md>), [streaming](<https://devfeed.tech/tags/streaming.md>), [voice-agents](<https://devfeed.tech/tags/voice-agents.md>)

### AI overview

This tutorial explains how to keep voice agents responsive while long-running backend tools run. It covers streaming progress updates, background execution, and response scheduling options in Google's Agent Development Kit (ADK), with Python examples for the Gemini Live API. In one benchmark sales-report workflow, switching from a blocking tool to a streaming tool reduced time to the agent's first words from 14.4 seconds to 1.2 seconds; backend execution time itself is unchanged.

### Source excerpt

The open-source Agent Development Kit (ADK) and gemini-3.8-live enable developers to build responsive voice agents by implementing non-blocking tools in Python. Developers can use streaming tools to yield progress updates or background execution patterns to keep conversations fluid during long-running backend queries. Additionally, configuring response scheduling behaviors such as WHEN_IDLE, INTERRUPT, or SILENT gives developers precise control over when and how tool results are delivered into the conversation.

## How Deepgram optimized voice AI at scale with AWS Trainium

DevFeed: [How Deepgram optimized voice AI at scale with AWS Trainium](<https://devfeed.tech/articles/how-deepgram-optimized-voice-ai-at-scale-with-aws-trainium-90433.md>)

Original publisher: [Read original article](<https://aws.amazon.com/blogs/machine-learning/how-deepgram-optimized-voice-ai-at-scale-with-aws-trainium/>)

Author: Pierre Lienhart

Published: 2026-10-19T16:23:22Z

Content type: article

Language: en

Sources: [Artificial Intelligence](<https://devfeed.tech/sources/artificial-intelligence.md>)

Topics: [AWS AI chips](<https://devfeed.tech/topics/aws-ai-chips.md>), [AWS Trainium](<https://devfeed.tech/topics/aws-trainium.md>), [neuron](<https://devfeed.tech/topics/neuron.md>), [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>), [data-engineering](<https://devfeed.tech/topics/data-engineering.md>), [chainguard vms](<https://devfeed.tech/topics/chainguard-vms.md>)

Tags: [advanced-300](<https://devfeed.tech/tags/advanced-300.md>), [ai](<https://devfeed.tech/tags/ai.md>), [aws](<https://devfeed.tech/tags/aws.md>), [aws-trainium](<https://devfeed.tech/tags/aws-trainium.md>), [customer-solutions](<https://devfeed.tech/tags/customer-solutions.md>), [deepgram](<https://devfeed.tech/tags/deepgram.md>), [inference](<https://devfeed.tech/tags/inference.md>), [neuron](<https://devfeed.tech/tags/neuron.md>), [parallel](<https://devfeed.tech/tags/parallel.md>), [rust](<https://devfeed.tech/tags/rust.md>), [speech-to-text](<https://devfeed.tech/tags/speech-to-text.md>), [throughput](<https://devfeed.tech/tags/throughput.md>), [trainium](<https://devfeed.tech/tags/trainium.md>), [voice](<https://devfeed.tech/tags/voice.md>), [voice-ai](<https://devfeed.tech/tags/voice-ai.md>)

### AI overview

Deepgram and AWS ported the Nova-3 speech-to-text model to Trainium2 for offline batch inference. They tuned kernels and built a Rust data-parallel runtime. On the evaluated workloads, the setup processed 11-70% more audio per dollar than Deepgram's previous compute options.

### Source excerpt

Deepgram worked with the AWS Forward Deployed Engineering team to port its Nova-3 speech-to-text model to AWS Trainium2 for offline batch inference. This post covers porting to the AWS Neuron stack, kernel-level tuning, a data-parallel Rust runtime, and the throughput and cost results on batch transcription.

## Meet Grok Voice 3.0: SpaceXAI Voice Agent Model Ranks #1 on Artificial Analysis

DevFeed: [Meet Grok Voice 3.0: SpaceXAI Voice Agent Model Ranks #1 on Artificial Analysis](<https://devfeed.tech/articles/meet-grok-voice-3-0-spacexai-voice-agent-model-ranks-1-on-artificial-analysis-87312.md>)

Original publisher: [Read original article](<https://www.marktechpost.com/2026/10/14/meet-grok-voice-3-0-spacexai-voice-agent-model-ranks-1-on-artificial-analysis/>)

Author: Michal Sutter

Published: 2026-10-15T05:18:54Z

Content type: news

Language: en

Sources: [MarkTechPost](<https://devfeed.tech/sources/marktechpost.md>)

Topics: [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>), [Conversational AI](<https://devfeed.tech/topics/conversational-ai.md>), [AI Integration Strategies](<https://devfeed.tech/topics/ai-integration-strategies.md>), [function calling](<https://devfeed.tech/topics/function-calling.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [audio-language-model](<https://devfeed.tech/tags/audio-language-model.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [collections](<https://devfeed.tech/tags/collections.md>), [editors-pick](<https://devfeed.tech/tags/editors-pick.md>), [grok](<https://devfeed.tech/tags/grok.md>), [language-model](<https://devfeed.tech/tags/language-model.md>), [model](<https://devfeed.tech/tags/model.md>), [new-releases](<https://devfeed.tech/tags/new-releases.md>), [python](<https://devfeed.tech/tags/python.md>), [spacexai](<https://devfeed.tech/tags/spacexai.md>), [staff](<https://devfeed.tech/tags/staff.md>), [tech-news](<https://devfeed.tech/tags/tech-news.md>), [technology](<https://devfeed.tech/tags/technology.md>), [tts](<https://devfeed.tech/tags/tts.md>), [uncategorized](<https://devfeed.tech/tags/uncategorized.md>), [voice-ai](<https://devfeed.tech/tags/voice-ai.md>)

### AI overview

SpaceXAI's Grok Voice 3.0 is a hosted speech-to-speech model exposed through a realtime API. It ranks first on the Artificial Analysis Speech-to-Speech Index with a score of 86.9, though it does not lead every component benchmark. The article describes its latency, pricing, API compatibility, audio options, and deployment details for developers.

### Source excerpt

SpaceXAI just released Grok Voice 3.0. Grok Voice 3.0 is a SpaceXAI speech-to-speech model that listens, reasons, calls tools and speaks. It now ranks first on the Artificial Analysis Speech-to-Speech Index with a score of 86.9. For developers, the endpoint, events and $0.08 per minute price are unchanged from Grok Voice 2.0. TL;DR What is [...] The post Meet Grok Voice 3.0: SpaceXAI Voice Agent Model Ranks #1 on Artificial Analysis appeared first on MarkTechPost.

## How to Build a Voice AI Agent with OpenAI's GPT-Live, Twilio Agent Connect, and TypeScript

DevFeed: [How to Build a Voice AI Agent with OpenAI's GPT-Live, Twilio Agent Connect, and TypeScript](<https://devfeed.tech/articles/how-to-build-a-voice-ai-agent-with-openai-s-gpt-live-twilio-agent-connect-and-typescript-88576.md>)

Original publisher: [Read original article](<https://www.twilio.com/en-us/blog/developers/tutorials/integrations/tac-gpt-live-voice-ai-agent-typescript>)

Author: Wen Zhu, Xinghao Huang

Published: 2026-10-15T00:00:00Z

Content type: tutorial

Language: en

Sources: [Twilio](<https://devfeed.tech/sources/twilio-blog.md>), [Twilio Blog](<https://devfeed.tech/sources/twilio-blog-2.md>)

Topics: [Conversational AI](<https://devfeed.tech/topics/conversational-ai.md>), [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>), [twilio api](<https://devfeed.tech/topics/twilio-api.md>), [TypeScript](<https://devfeed.tech/topics/typescript.md>), [ChatGPT API](<https://devfeed.tech/topics/chatgpt-api.md>), [OpenAI](<https://devfeed.tech/topics/openai.md>), [vercel ai sdk](<https://devfeed.tech/topics/vercel-ai-sdk.md>), [twilio programmable video](<https://devfeed.tech/topics/twilio-programmable-video.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [api](<https://devfeed.tech/tags/api.md>), [connect](<https://devfeed.tech/tags/connect.md>), [developer-insights](<https://devfeed.tech/tags/developer-insights.md>), [how-to](<https://devfeed.tech/tags/how-to.md>), [openai](<https://devfeed.tech/tags/openai.md>), [twilio](<https://devfeed.tech/tags/twilio.md>), [typescript](<https://devfeed.tech/tags/typescript.md>), [voice](<https://devfeed.tech/tags/voice.md>), [voice-ai](<https://devfeed.tech/tags/voice-ai.md>)

### AI overview

This tutorial walks through building inbound and outbound phone agents that use OpenAI's GPT-Live for speech-to-speech conversations and Twilio Agent Connect for call handling. It demonstrates TypeScript setup, Responses API tool delegation, and configuring Twilio voice calls.

### Source excerpt

Build inbound and outbound voice AI assistants that can handle tool calls using TypeScript, Twilio Agent Connect, and OpenAI's GPT-Live.

## Grok Voice 3.0 from SpaceXSI outscores OpenAI and Google on speech-to-speech index

DevFeed: [Grok Voice 3.0 from SpaceXSI outscores OpenAI and Google on speech-to-speech index](<https://devfeed.tech/articles/grok-voice-3-0-from-spacexsi-outscores-openai-and-google-on-speech-to-speech-index-87334.md>)

Original publisher: [Read original article](<https://siliconangle.com/2026/10/14/grok-voice-3-0-from-spacexsi-outscores-openai-and-google-on-speech-to-speech-index/>)

Author: Duncan Riley

Published: 2026-10-14T22:27:54Z

Content type: news

Language: en

Sources: [SiliconANGLE](<https://devfeed.tech/sources/siliconangle.md>)

Topics: [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>), [Speech Processing](<https://devfeed.tech/topics/speech-processing.md>), [OpenAI](<https://devfeed.tech/topics/openai.md>), [ChatGPT API](<https://devfeed.tech/topics/chatgpt-api.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-benchmarks](<https://devfeed.tech/tags/ai-benchmarks.md>), [api](<https://devfeed.tech/tags/api.md>), [artificial-analysis](<https://devfeed.tech/tags/artificial-analysis.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [big-bench-audio](<https://devfeed.tech/tags/big-bench-audio.md>), [customer-support](<https://devfeed.tech/tags/customer-support.md>), [elon-musk](<https://devfeed.tech/tags/elon-musk.md>), [eva-bench](<https://devfeed.tech/tags/eva-bench.md>), [full-duplex-bench](<https://devfeed.tech/tags/full-duplex-bench.md>), [gemini-3-8-live](<https://devfeed.tech/tags/gemini-3-8-live.md>), [google](<https://devfeed.tech/tags/google.md>), [gpt-6-astra](<https://devfeed.tech/tags/gpt-6-astra.md>), [gpt-live-1](<https://devfeed.tech/tags/gpt-live-1.md>), [gpt-realtime-2-1](<https://devfeed.tech/tags/gpt-realtime-2-1.md>), [grok](<https://devfeed.tech/tags/grok.md>), [grok-bot](<https://devfeed.tech/tags/grok-bot.md>), [grok-voice-2-0](<https://devfeed.tech/tags/grok-voice-2-0.md>), [grok-voice-3-0](<https://devfeed.tech/tags/grok-voice-3-0.md>), [index](<https://devfeed.tech/tags/index.md>), [model](<https://devfeed.tech/tags/model.md>), [models](<https://devfeed.tech/tags/models.md>), [news](<https://devfeed.tech/tags/news.md>), [nitin-aggarwal](<https://devfeed.tech/tags/nitin-aggarwal.md>), [openai](<https://devfeed.tech/tags/openai.md>), [openai-group](<https://devfeed.tech/tags/openai-group.md>), [servicenow](<https://devfeed.tech/tags/servicenow.md>), [siliconangle](<https://devfeed.tech/tags/siliconangle.md>), [spacex](<https://devfeed.tech/tags/spacex.md>), [spacexai](<https://devfeed.tech/tags/spacexai.md>), [spacexsi](<https://devfeed.tech/tags/spacexsi.md>), [speech-to-speech](<https://devfeed.tech/tags/speech-to-speech.md>), [speech-to-speech-index](<https://devfeed.tech/tags/speech-to-speech-index.md>), [starlink](<https://devfeed.tech/tags/starlink.md>), [tau-voice](<https://devfeed.tech/tags/tau-voice.md>), [tesla](<https://devfeed.tech/tags/tesla.md>), [the-latest](<https://devfeed.tech/tags/the-latest.md>), [top-story-6](<https://devfeed.tech/tags/top-story-6.md>), [voice](<https://devfeed.tech/tags/voice.md>), [voice-agent-api](<https://devfeed.tech/tags/voice-agent-api.md>), [voice-agents](<https://devfeed.tech/tags/voice-agents.md>), [xai](<https://devfeed.tech/tags/xai.md>)

### AI overview

SpaceXSI released Grok Voice 3.0, a speech-to-speech model for customer support and sales voice agents. Artificial Analysis ranked it first on its Speech-to-Speech Index, while the article reports different results across other benchmarks, including a narrow OpenAI lead on Tau-Voice. The model is available through the Voice Agent API, and existing users of the grok-voice-latest alias were moved to it automatically.

### Source excerpt

Elon Musk's artificial intelligence company SpaceXSI, formerly xAI, today released Grok Voice 3.0, a new version of the speech-to-speech model behind its Voice Agent API. SpaceXSI is pitching the model to businesses that handle customer support and sales calls with voice agents. Independent testing firm Artificial Analysis ranks it first among models of its type. Artificial [...] The post Grok Voice 3.0 from SpaceXSI outscores OpenAI and Google on speech-to-speech index appeared first on SiliconANGLE.

## Open-Source Voice AI runs entirely on ESP32-S3

DevFeed: [Open-Source Voice AI runs entirely on ESP32-S3](<https://devfeed.tech/articles/open-source-voice-ai-runs-entirely-on-esp32-s3-85404.md>)

Original publisher: [Read original article](<https://blog.adafruit.com/2026/10/12/open-source-voice-ai-runs-entirely-on-esp32-s3/>)

Author: Anne Barela

Published: 2026-10-12T18:44:04Z

Content type: news

Language: en

Sources: [Adafruit Industries](<https://devfeed.tech/sources/adafruit-industries-makers-hackers-artists-designers-and-engineers.md>)

Topics: [Toit](<https://devfeed.tech/topics/toit.md>), [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>), [lunar foundation model](<https://devfeed.tech/topics/lunar-foundation-model.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [dataset](<https://devfeed.tech/tags/dataset.md>), [esp32-s3](<https://devfeed.tech/tags/esp32-s3.md>), [espressif](<https://devfeed.tech/tags/espressif.md>), [firmware](<https://devfeed.tech/tags/firmware.md>), [gguf](<https://devfeed.tech/tags/gguf.md>), [inference](<https://devfeed.tech/tags/inference.md>), [model-inference](<https://devfeed.tech/tags/model-inference.md>), [speech](<https://devfeed.tech/tags/speech.md>), [speech-recognition](<https://devfeed.tech/tags/speech-recognition.md>), [speech-synthesis](<https://devfeed.tech/tags/speech-synthesis.md>)

### AI overview

Ivy AI is an open-source voice assistant that runs speech recognition, language model inference and speech synthesis locally on an ESP32-S3, using Wi-Fi only for time synchronization. The article describes its hardware, four-bit model quantization, performance, memory constraints and licensing considerations.

### Source excerpt

Ananthu Ashok, writing for Open Source For You, covers Ivy AI, a voice assistant from Anankush Pandit that runs speech recognition, language model inference and speech synthesis on an ESP32-S3. Firmware, model conversion tools, host-side tests and documentation are published under GPL-3.0-or-later. WiFi is used only for time sync, so the voice pipeline works without [...]

## Top 12 AI Live Translation Tools in 2026: Latency, Languages, Pricing

DevFeed: [Top 12 AI Live Translation Tools in 2026: Latency, Languages, Pricing](<https://devfeed.tech/articles/top-12-ai-live-translation-tools-in-2026-latency-languages-pricing-85255.md>)

Original publisher: [Read original article](<https://www.marktechpost.com/2026/10/12/top-12-ai-live-translation-tools-2026/>)

Author: Michal Sutter

Published: 2026-10-12T08:20:49Z

Content type: comparison

Language: en

Sources: [MarkTechPost](<https://devfeed.tech/sources/marktechpost.md>)

Topics: [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>), [machine-translation](<https://devfeed.tech/topics/machine-translation.md>), [Whisper](<https://devfeed.tech/topics/whisper.md>), [deepl](<https://devfeed.tech/topics/deepl.md>)

Tags: [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [ai](<https://devfeed.tech/tags/ai.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [audio-language-model](<https://devfeed.tech/tags/audio-language-model.md>), [editors-pick](<https://devfeed.tech/tags/editors-pick.md>), [for-devs](<https://devfeed.tech/tags/for-devs.md>), [latency](<https://devfeed.tech/tags/latency.md>), [pricing](<https://devfeed.tech/tags/pricing.md>), [staff](<https://devfeed.tech/tags/staff.md>), [tech-news](<https://devfeed.tech/tags/tech-news.md>), [technology](<https://devfeed.tech/tags/technology.md>), [top](<https://devfeed.tech/tags/top.md>), [translation](<https://devfeed.tech/tags/translation.md>), [voice-ai](<https://devfeed.tech/tags/voice-ai.md>)

### AI overview

This comparison reviews 12 AI live translation tools for conversations, meetings, events and webinars, covering language support, latency, platforms and pricing. It notes that language counts are not directly comparable and recommends piloting with your own language pairs.

### Source excerpt

We checked 12 AI live translation tools, including Palabra.ai, Google Live Translate, DeepL Voice, Teams Interpreter and Wordly, against official pages for languages, latency, platforms and pricing. The post Top 12 AI Live Translation Tools in 2026: Latency, Languages, Pricing appeared first on MarkTechPost.

## These execs think voice AI hasn't reached its ChatGPT moment yet

DevFeed: [These execs think voice AI hasn't reached its ChatGPT moment yet](<https://devfeed.tech/articles/these-execs-think-voice-ai-hasn-t-reached-its-chatgpt-moment-yet-84894.md>)

Original publisher: [Read original article](<https://techcrunch.com/2026/10/11/these-execs-think-voice-ai-hasnt-reached-its-chatgpt-moment-yet/>)

Author: Ivan Mehta

Published: 2026-10-11T14:00:00Z

Content type: news

Language: en

Sources: [TechCrunch | AI News & Artificial Intelligence](<https://devfeed.tech/sources/techcrunch-ai-news-artificial-intelligence.md>)

Topics: [Conversational AI](<https://devfeed.tech/topics/conversational-ai.md>), [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>), [Model Routing](<https://devfeed.tech/topics/model-routing.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [asr](<https://devfeed.tech/tags/asr.md>), [capture](<https://devfeed.tech/tags/capture.md>), [customer-service](<https://devfeed.tech/tags/customer-service.md>), [interface](<https://devfeed.tech/tags/interface.md>), [otter](<https://devfeed.tech/tags/otter.md>), [voice-ai](<https://devfeed.tech/tags/voice-ai.md>)

### AI overview

Executives from PolyAI and Otter say voice AI has yet to reach a ChatGPT-like turning point. They point to faster reasoning, more accurate speech recognition and transcription, natural conversation, and transparency about recording and AI use as challenges affecting user confidence and trust.

### Source excerpt

Voice AI's often misses important points for its context layer, and causes the whole pipeline to break

## Building a Multi-Channel AI Agent with Twilio Conversations and eve

DevFeed: [Building a Multi-Channel AI Agent with Twilio Conversations and eve](<https://devfeed.tech/articles/building-a-multi-channel-ai-agent-with-twilio-conversations-and-eve-83868.md>)

Original publisher: [Read original article](<https://www.twilio.com/en-us/blog/developers/building-agent-twilio-conversations-eve>)

Author: Marius Obert

Published: 2026-10-08T00:00:00Z

Content type: tutorial

Language: en

Sources: [Twilio](<https://devfeed.tech/sources/twilio-blog.md>), [Twilio Blog](<https://devfeed.tech/sources/twilio-blog-2.md>)

Topics: [twilio api](<https://devfeed.tech/topics/twilio-api.md>), [twilio programmable video](<https://devfeed.tech/topics/twilio-programmable-video.md>), [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>), [AI Chat](<https://devfeed.tech/topics/ai-chat.md>), [Microsoft Agent Framework](<https://devfeed.tech/topics/microsoft-agent-framework.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [ai-agent-memory](<https://devfeed.tech/tags/ai-agent-memory.md>), [ai-sdk](<https://devfeed.tech/tags/ai-sdk.md>), [api](<https://devfeed.tech/tags/api.md>), [conversation-intelligence](<https://devfeed.tech/tags/conversation-intelligence.md>), [conversation-orchestrator](<https://devfeed.tech/tags/conversation-orchestrator.md>), [conversational-ai](<https://devfeed.tech/tags/conversational-ai.md>), [cross-channel-messaging](<https://devfeed.tech/tags/cross-channel-messaging.md>), [customer-memory](<https://devfeed.tech/tags/customer-memory.md>), [developer-insights](<https://devfeed.tech/tags/developer-insights.md>), [eve-framework](<https://devfeed.tech/tags/eve-framework.md>), [how-to](<https://devfeed.tech/tags/how-to.md>), [llm](<https://devfeed.tech/tags/llm.md>), [multi-channel-ai-agent](<https://devfeed.tech/tags/multi-channel-ai-agent.md>), [sms-and-whatsapp](<https://devfeed.tech/tags/sms-and-whatsapp.md>), [twilio](<https://devfeed.tech/tags/twilio.md>), [twilio-conversations](<https://devfeed.tech/tags/twilio-conversations.md>), [twilio-memory-store](<https://devfeed.tech/tags/twilio-memory-store.md>), [typescript](<https://devfeed.tech/tags/typescript.md>), [vercel](<https://devfeed.tech/tags/vercel.md>), [vercel-eve](<https://devfeed.tech/tags/vercel-eve.md>), [whatsapp](<https://devfeed.tech/tags/whatsapp.md>)

### AI overview

This tutorial shows how to combine Vercel's eve framework with Twilio Conversations to build an AI agent that keeps customer context across SMS and WhatsApp. It separates short-term conversation transcripts from persistent customer memory, then walks through webhook routing, cross-channel identity linking, and injecting stored observations into model prompts.

### Source excerpt

Learn how to build a multi-channel AI agent with Vercel's eve and Twilio Conversations. Unify SMS and WhatsApp into one thread, give your agent long-term customer memory with Twilio Memory Store, and swap the LLM with one line of TypeScript.

## ElevenLabs is now on OpenRouter

DevFeed: [ElevenLabs is now on OpenRouter](<https://devfeed.tech/articles/elevenlabs-is-now-on-openrouter-78539.md>)

Original publisher: [Read original article](<https://openrouter.ai/blog/announcements/elevenlabs-on-openrouter/>)

Author: Sasha Blumenfeld

Published: 2026-10-07T00:00:00Z

Content type: tutorial

Language: en

Sources: [OpenRouter](<https://devfeed.tech/sources/openrouter.md>)

Topics: [asr](<https://devfeed.tech/topics/asr.md>), [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>), [Conversational AI](<https://devfeed.tech/topics/conversational-ai.md>), [ChatGPT API](<https://devfeed.tech/topics/chatgpt-api.md>)

Tags: [agents](<https://devfeed.tech/tags/agents.md>), [ai-api](<https://devfeed.tech/tags/ai-api.md>), [elevenlabs](<https://devfeed.tech/tags/elevenlabs.md>), [latency](<https://devfeed.tech/tags/latency.md>), [low-latency](<https://devfeed.tech/tags/low-latency.md>), [meetings](<https://devfeed.tech/tags/meetings.md>), [openrouter](<https://devfeed.tech/tags/openrouter.md>), [speech-to-text](<https://devfeed.tech/tags/speech-to-text.md>), [text-to-speech](<https://devfeed.tech/tags/text-to-speech.md>)

### AI overview

OpenRouter makes ElevenLabs speech models available through its API, allowing applications to add text-to-speech and speech-to-text features using an OpenRouter API key. The article walks through narration, a voice agent that transcribes questions and speaks responses, and meeting transcription with speaker labels. It also covers model selection, request parameters, limits, and pricing.

### Source excerpt

ElevenLabs is now on OpenRouter with 9 Text to Speech and 2 Speech to Text models. Deliver speech with human-like emotion with Eleven v4, give your agent low-latency replies with v4 Turbo, and turn meetings into notes that show who said what with Scribe v2. Here's which model to pick, and one project built start to finish.

## Designing a Low-Latency Speech-to-Text Pipeline for Dictation

DevFeed: [Designing a Low-Latency Speech-to-Text Pipeline for Dictation](<https://devfeed.tech/articles/designing-a-low-latency-speech-to-text-pipeline-for-dictation-64954.md>)

Original publisher: [Read original article](<https://hackernoon.com/designing-a-low-latency-speech-to-text-pipeline-for-dictation?source=rss>)

Author: AssemblyAI

Published: 2026-10-05T12:37:36Z

Content type: article

Language: en

Sources: [HackerNoon](<https://devfeed.tech/sources/hackernoon.md>)

Topics: [asr](<https://devfeed.tech/topics/asr.md>), [WebSocket](<https://devfeed.tech/topics/websocket.md>), [Streaming](<https://devfeed.tech/topics/streaming.md>), [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>)

Tags: [api](<https://devfeed.tech/tags/api.md>), [assemblyai](<https://devfeed.tech/tags/assemblyai.md>), [cli](<https://devfeed.tech/tags/cli.md>), [code](<https://devfeed.tech/tags/code.md>), [dictation-api](<https://devfeed.tech/tags/dictation-api.md>), [good-company](<https://devfeed.tech/tags/good-company.md>), [latency](<https://devfeed.tech/tags/latency.md>), [pipeline](<https://devfeed.tech/tags/pipeline.md>), [speech](<https://devfeed.tech/tags/speech.md>), [speech-recognition](<https://devfeed.tech/tags/speech-recognition.md>), [speech-to-text](<https://devfeed.tech/tags/speech-to-text.md>), [streaming](<https://devfeed.tech/tags/streaming.md>), [streaming-transcription](<https://devfeed.tech/tags/streaming-transcription.md>), [voice-interfaces](<https://devfeed.tech/tags/voice-interfaces.md>), [web-speech-api](<https://devfeed.tech/tags/web-speech-api.md>), [websocket](<https://devfeed.tech/tags/websocket.md>)

### AI overview

The article compares browser-based, streaming, asynchronous, and synchronous approaches to push-to-talk dictation. It argues that a synchronous request suits short utterances, then describes ways to reduce latency, handle transcription and cleanup separately, guard against spoken prompt injection, and weigh API costs. It also discusses a reference implementation and the limits of turning speech into polished writing.

### Source excerpt

Push-to-talk dictation doesn't need a WebSocket. Here's how streaming, async, and a sync dictation API compare on latency, billing, and failure modes.

## Microsoft targets ultra-realistic voice agents with its first streaming transcription model

DevFeed: [Microsoft targets ultra-realistic voice agents with its first streaming transcription model](<https://devfeed.tech/articles/microsoft-targets-ultra-realistic-voice-agents-with-its-first-streaming-transcription-model-63645.md>)

Original publisher: [Read original article](<https://siliconangle.com/2026/10/01/microsoft-targets-ultra-realistic-voice-agents-with-its-first-streaming-transcription-model/>)

Author: Mike Wheatley

Published: 2026-10-02T01:31:31Z

Content type: news

Language: en

Sources: [SiliconANGLE](<https://devfeed.tech/sources/siliconangle.md>)

Topics: [Whisper](<https://devfeed.tech/topics/whisper.md>), [Conversational AI](<https://devfeed.tech/topics/conversational-ai.md>), [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>), [Streaming](<https://devfeed.tech/topics/streaming.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-transcription](<https://devfeed.tech/tags/ai-transcription.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [mai-models](<https://devfeed.tech/tags/mai-models.md>), [mai-thinking-1](<https://devfeed.tech/tags/mai-thinking-1.md>), [mai-transcribe-2-streaming](<https://devfeed.tech/tags/mai-transcribe-2-streaming.md>), [mai-voice-2-1](<https://devfeed.tech/tags/mai-voice-2-1.md>), [mai-voice-2-1-flash](<https://devfeed.tech/tags/mai-voice-2-1-flash.md>), [microsoft](<https://devfeed.tech/tags/microsoft.md>), [microsoft-ai](<https://devfeed.tech/tags/microsoft-ai.md>), [mustafa-suleyman](<https://devfeed.tech/tags/mustafa-suleyman.md>), [news](<https://devfeed.tech/tags/news.md>), [streaming-model](<https://devfeed.tech/tags/streaming-model.md>), [streaming-transcription](<https://devfeed.tech/tags/streaming-transcription.md>), [text-to-speech](<https://devfeed.tech/tags/text-to-speech.md>), [the-latest](<https://devfeed.tech/tags/the-latest.md>), [transcription](<https://devfeed.tech/tags/transcription.md>), [voice-agents](<https://devfeed.tech/tags/voice-agents.md>)

### AI overview

Microsoft released MAI-Transcribe-2-Streaming, its first streaming transcription model, alongside two text-to-speech models. The transcription model accepts speech over WebSocket, updates transcripts as a person speaks, supports more than 60 languages, and is listed at 54 cents per audio hour. Microsoft says its first transcript hypotheses arrive in 320 milliseconds on average, with performance depending on network and reply-generation conditions. The releases target developers building voice agents, and the article describes how combining transcription, reasoning, and speech generation gives developers control over quality, latency, and cost.

### Source excerpt

Microsoft Corp. today expanded its MAI artificial intelligence model family with its first streaming transcription model, debuting alongside two others focused on text-to-speech. They're designed for developers who want to build voice agents that can listen to people's voices and reply instantly, similar to how humans talk to one another. The most consequential of the [...] The post Microsoft targets ultra-realistic voice agents with its first streaming transcription model appeared first on SiliconANGLE.

## DIY AI voice assistant with ESP32 and animated OLED eyes

DevFeed: [DIY AI voice assistant with ESP32 and animated OLED eyes](<https://devfeed.tech/articles/diy-ai-voice-assistant-with-esp32-and-animated-oled-eyes-62642.md>)

Original publisher: [Read original article](<https://blog.adafruit.com/2026/09/30/diy-ai-voice-assistant-with-esp32-and-animated-oled-eyes/>)

Author: Anne Barela

Published: 2026-09-30T20:37:33Z

Content type: article

Language: en

Sources: [Adafruit Industries](<https://devfeed.tech/sources/adafruit-industries-makers-hackers-artists-designers-and-engineers.md>)

Topics: [ESP32](<https://devfeed.tech/topics/esp32.md>), [speech-to-speech](<https://devfeed.tech/topics/speech-to-speech.md>), [Whisper](<https://devfeed.tech/topics/whisper.md>), [ChatGPT](<https://devfeed.tech/topics/chatgpt.md>)

Tags: [adafruit](<https://devfeed.tech/tags/adafruit.md>), [ai](<https://devfeed.tech/tags/ai.md>), [arduino](<https://devfeed.tech/tags/arduino.md>), [arduino-ide](<https://devfeed.tech/tags/arduino-ide.md>), [audio](<https://devfeed.tech/tags/audio.md>), [blocking](<https://devfeed.tech/tags/blocking.md>), [displays](<https://devfeed.tech/tags/displays.md>), [esp32](<https://devfeed.tech/tags/esp32.md>), [espressif](<https://devfeed.tech/tags/espressif.md>), [freertos](<https://devfeed.tech/tags/freertos.md>), [oled](<https://devfeed.tech/tags/oled.md>), [sound](<https://devfeed.tech/tags/sound.md>), [speech](<https://devfeed.tech/tags/speech.md>), [speech-to-text](<https://devfeed.tech/tags/speech-to-text.md>), [text-to-speech](<https://devfeed.tech/tags/text-to-speech.md>), [voice](<https://devfeed.tech/tags/voice.md>), [voice-assistant](<https://devfeed.tech/tags/voice-assistant.md>), [whisper](<https://devfeed.tech/tags/whisper.md>)

### AI overview

This DIY voice assistant runs on an ESP32: it sends recorded speech to the OpenAI Whisper API for transcription, uses GPT to generate a reply, and streams text-to-speech audio to a speaker. Animated OLED eyes run in a separate FreeRTOS task from the audio and network pipeline, so blocking HTTP calls do not stall the display.

### Source excerpt

oeadmin at Open Electronics covers a voice assistant project by jayesh_nawani that runs entirely on an ESP32. An I2S microphone picks up speech, the clip is packaged as a WAV file and sent to the OpenAI Whisper API for transcription. GPT writes a reply and the TTS audio streams back out through a MAX98357A amplifier [...]

[Next page](<https://devfeed.tech/topics/speech-to-speech.md?cursor=WyIyMDI2LTA5LTMwVDIwOjM3OjMzKzAwOjAwIiwgIjY1NDBkODgzLWRkN2ItNDRkMy1iNDFmLTU2MDMwNTI3MzE2YSJd>)