# higurashi-ai/Haru

> Inference code for Haru-4B, a multilingual full-duplex spoken dialogue model that listens and speaks at the same time, like on a phone call. Includes PyTorch, MLX and vLLM-Omni implementations plus a web demo.

- Magnitude: 5.3 out of 10 — Early signal
- Stars: 248 total · +35 stars measured 2026-11-14, 00:03–11:13 UTC, ≈ 93 by evening
- Star trust: star growth looks organic
- Category: Generative media · Language: Python · License: Apache-2.0 · Created: 2026-11-12 · Last push: 2026-11-13
- GitHub: https://github.com/higurashi-ai/Haru · Homepage: https://huggingface.co/higurashi-ai/Haru-4B · Page: https://gitnova.dev/en/r/higurashi-ai/Haru

## Useful for

- Run the web demo on a GPU and talk to the model with a chosen character and voice
- Build the native macOS app on Apple silicon via haru_mlx
- Serve many concurrent conversations through the vLLM-Omni plugin

## How to get started

```sh
pip install -e ./haru
python -m haru.server --hf-repo higurashi-ai/Haru-4B
```

Copied as is from the “quick start” section of the README. GitNova has not run these commands: check them in the original before running: https://github.com/higurashi-ai/Haru#readme

## Why it’s here

- Star-counter measurements on 2026-11-14 (UTC), 00:03–11:13: 213 → 248 stars (+35). This is the change over that interval.
- Estimated end-of-day forecast: about +93 stars, using observed gains and the previous day.
- The repository is 2 days old and already has 248 stars. With less than two weeks of history, there's no usual pace to compare the spike against yet.
- About 9 forks a day — people are taking the code.

## Star trust

Star growth looks organic.

Star-trust labels are heuristics based on the repository’s behavior, not a check of every stargazer.

## Numbers

- Forks: 14
- Issues and pull requests: 3
- Watchers: 4
- Average over the last week: 109 per day
- Usual pace: too little history (under two weeks)
- Stars in the last hour (measured): 2

## Stars per day, last 3 days (oldest → newest, today is partial)

2026-11-12 … 2026-11-14: 92, 143, 35

## Magnitude by day, last 2 days

11-13 5.3, 11-14 5.6

## Similar by description

1. **OpenMOSS/MOSS-Live** — 1.8 · Steady · AI agents · +6 stars measured 2026-11-14, 00:09–11:17 UTC, ≈ 12 by evening
   MOSS-Live is a real-time multimodal conversational system that listens and speaks simultaneously, handles interruptions, understands live camera and screen input, and runs background tasks without pausing the conversation.
   Full card: https://gitnova.dev/en/r/OpenMOSS/MOSS-Live.md

---

Magnitude (0–10) measures how fast and how unusually interest in a repository is growing right now. It is not a quality score.
Days are UTC. “So far today” is a fact; “expected by the end of the day” is a forecast.
Summaries and use cases are written by an LLM (DeepSeek V4.1 Flash) from the README and may be inaccurate: verify specific claims (benchmarks, speed, hardware) in the repository itself.
Data as of 2026-11-14 11:22 UTC, updated every 30 minutes.