# Standard-Intelligence/hertz-dev

first base model for full-duplex conversational audio

Repository: https://github.com/Standard-Intelligence/hertz-dev
Canonical: https://ross.abutalabs.com/products/hertz-dev
Homepage: https://si.inc
Language: Python
License: Apache-2.0
License Family: permissive
Last push: 2025-01-05T23:43:25+00:00

## Health v2 (maintenance only)
Score: 22/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 47
- inputs: {"age_days": 668, "days_push": 605, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1799, forks 114 (observed 2026-08-28T04:05:37.851877+00:00)

## What it is
Hertz-dev is an open-source 8.5B parameter autoregressive base model for full-duplex conversational audio, released by Standard Intelligence. It includes inference scripts for generating one- or two-channel audio completions and live voice chat via a client-server setup or a browser-based WebRTC client.

## Use cases
- run full-duplex voice conversations with an AI model
- generate conversational speech audio from a prompt
- build a live microphone voice chatbot
- experiment with speech generation base models
- host a browser-based voice chat client with WebRTC

## When to choose
- you need a full-duplex conversational audio model you can self-host
- you want to research or fine-tune speech generation base models
- you have CUDA GPUs and want live voice interaction with an open model

## When to avoid
- you need production-ready, polished voice assistant software
- you lack a CUDA GPU with sufficient memory for an 8.5B parameter model
- you need text-only chat or non-speech audio generation

## Facets
- artifact type: library
- maturity: experimental
- function: speech-recognition, machine-learning, llm-inference, audio-processing
- domain: speech-processing, large-language-models, artificial-intelligence
- platform: python
- tags: full-duplex-audio, speech-generation, conversational-ai, autoregressive-model, webrtc, voice-chat, 8.5b-parameters, audio, linux, macos, gpu, web-server

## Member repositories
- Standard-Intelligence/hertz-dev (main) score 22

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:37.851877+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:22:16.482276+00:00, confidence not recorded.
  - readme: https://github.com/Standard-Intelligence/hertz-dev (fetched 2026-08-28T04:05:37.851877+00:00, sha 9441345914bd)
  - homepage: https://si.inc (fetched 2026-08-29T11:01:08.305242+00:00, sha d93f7b7398e9)
- Data as of 2026-08-30T08:39:29.467469+00:00.