Use cases

Public projects that use otoSpeech, each linked to its own paper, model card, or code. Last checked Nov 4, 2026.

Logos show the organizations of each project's authors or publishers. They do not mean product adoption, production use, or endorsement of oto.

Research

• ProjectOrganizationDataDate

• TurnBench

  Baseline turn-taking models (VAP, Mimi, WavLM) trained on otoSpeech. The evaluation audio is separate data from Mundo AI.

  
  Turn 104h Aug 25, 2026 Source ↗

• DualTurn

  A turn-taking model trained on otoSpeech, published with its weights and derived data.

  
  280h Mar 9, 2026 Source ↗

• Mind the Gap

  500 hours of synthetic conversation built from otoSpeech to train multi-talker speech recognition.

  
  141h May 14, 2026 Source ↗

• TurnMoshi

  A turn-taking model trained and distilled on about 100 hours of otoSpeech.

  
  Turn 104h Oct 3, 2026 Source ↗

• DuplexAct-Bench

  otoSpeech as evaluation data for natural backchannels.

  
  280h Sep 30, 2026 Source ↗

• TurnNat and SPEARBench

  Fine-tunes the weights of DualTurn, a model trained on otoSpeech, then reuses them as an evaluator.

  
  280h through DualTurn Jul 1, 2026 Source ↗

• MultiPartyTurnBench

  Evaluates DualTurn and TurnBench models, both trained on otoSpeech, on multi-party conversation.

  
  Models trained on otoSpeech Oct 14, 2026 Source ↗

• TACT, PALS, and TRIAD

  Three studies that use otoSpeech for timing evaluation, robustness training, and audits.

  People Make Things
  141h Sep 23, 2026 Source ↗

Models and products

• ProjectOrganizationDataDate

• Backchannel detector

  A streaming model that tells backchannels from interruptions, trained and evaluated in part on otoSpeech. The weights are public.

  Pipecat
  280h Oct 28, 2026 Source ↗

• Voice Reader

  An experimental model for tone of voice and sarcasm, with otoSpeech among its training and evaluation data.

  Pipecat
  otoSpeech Nov 2, 2026 Source ↗

• BarnOwl

  An interruption model that lists otoSpeech in its training data.

  Joveo
  280h, Task v2 Nov 2, 2026 Source ↗

• DuplexJev

  End-of-speech models fine-tuned on otoSpeech audio with labels derived from DualTurn.

  Adventists.ai
  141h Oct 13, 2026 Source ↗

• DualTurn Endpointing

  A small end-of-turn model trained on otoSpeech.

  
  otoSpeech Jun 18, 2026 Source ↗

• OpenRealtime

  An open realtime voice project that integrates DualTurn, a model trained on otoSpeech, and evaluates it offline.

  OpenRealtime
  280h through DualTurn Sep 22, 2026 Source ↗

Benchmarks and tools

• ProjectOrganizationDataDate

• FrontierML

  An end-of-turn benchmark with published trials that train on otoSpeech and on models derived from it.

  
  Turn 104h, 280h through DualTurn Oct 13, 2026 Source ↗

• eot-bench

  Evaluates a VAP model trained on otoSpeech across 14 languages.

  LiveKit
  Turn 104h through TurnBench Sep 14, 2026 Source ↗

• VAP-OTO

  A VAP model trained on otoSpeech, converted to Core ML and ONNX and added as an optional model in its speech SDKs.

  Soniqo
  Turn 104h through TurnBench Oct 30, 2026 Source ↗

• Lhotse

  A recipe for downloading and preparing otoSpeech in a speech data toolkit.

  Lhotse community
  141h Feb 26, 2026 Source ↗

• Multimodal data

  Scripts that download otoSpeech and convert it for training.

  Swiss AI
  141h Apr 30, 2026 Source ↗

otoSpeech is one of the largest speech AI datasets of our time, used in frontier model training and research at Sesame, NVIDIA, Pipecat, LiveKit, and Carnegie Mellon University.