Tazo
Audio-native decision model for live voice agents that picks respond, wait, yield, backchannel or escalate straight from call audio.
This summary is based on how curated lists describe the project; its README was not available to review.
Overview
Tazo is a decision model from Tazo Labs for live voice agents. According to the list summary, it reads call audio directly, without a transcript step, and returns a probability for each option the developer defines, such as respond, wait, yield, backchannel or escalate. It is offered as an API with access by request.
The company has published a write-up comparing Tazo with Jev, which works on text. No README or code is available in the directory, so details such as latency, accuracy or pricing are not recorded here.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Choice | What should the voice agent do now? | respond · wait · yield · backchannel · escalate |
How it works
- Stream live call audio to the model.
- The model scores the developer-defined turn-taking options.
- The voice agent acts on the highest-probability option.
Strengths
- No transcription step, per the list summary.
- Developer-defined options with probabilities.
Caveats
- API access by request only.
- No public code or published metrics in the directory.
At a glance
- Type
- Service
- Runtime
- hosted API
Use cases and integrations
How curated lists describe it
Tazo Labs' audio-native decision model for live voice agents: it reads the call audio with no transcript step and returns a probability for each option you define, such as respond, wait, yield, backchannel, or escalate. API access by request. Write-up: [Tazo vs Jev](https://tazolabs.com/blog/tazo-vs-jev-audio-decision-model).
AnotiaWang/awesome-decision-models
Frequently asked questions
What does Tazo decide?
Turn-taking actions for a voice agent, such as whether to respond, wait, yield, backchannel or escalate.
How is it different from Jev?
It reads audio directly instead of text; Tazo Labs has a write-up comparing the two.
How do I get access?
API access is by request through Tazo Labs.
Related projects
macOS voice agent where Jev picks each step's operation and target control, escalating to an LLM agent when unsure.
ArticleproductionToolkit for low-latency voice avatars where Jev can serve as a gate that judges turn ends and turn taking from speech transcripts.
ProjectproductionPythonCross-platform dictation app where a decision model chooses whether a personal-dictionary replacement actually fits the sentence.
ProjectbetaPythonPrompt reviewer for Claude Code and Codex where Jev decides whether each prompt needs review before a light model lists blind spots.
CLIbetaTypeScriptLiquid Glass voice typing for macOS 26 with an optional Jev fast path that types clean short sentences without waiting for LLM cleanup.
ProjectbetaSwiftChrome extension and web app that find YouTube sponsor reads with Jev from transcripts or audio and skip them.
PluginbetaJavaScript