# NVIDIA/personaplex

PersonaPlex code.

Repository: https://github.com/NVIDIA/personaplex
Canonical: https://ross.abutalabs.com/products/personaplex
Language: Python
License: MIT
License Family: permissive
Last push: 2026-03-02T16:07:10+00:00

## Health v2 (maintenance only)
Score: 47/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 70, release rhythm 35, longevity 17
- inputs: {"age_days": 240, "days_push": 184, "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 10393, forks 1449 (observed 2026-08-28T04:10:41.901938+00:00)

## What it is
PersonaPlex is a real-time, full-duplex speech-to-speech conversational model from NVIDIA that supports persona control via text role prompts and audio voice conditioning. It is built on the Moshi architecture and ships with a Python library, offline evaluation scripts, and a web server for live interaction.

## Use cases
- build a real-time voice chatbot with a consistent persona
- run full-duplex speech-to-speech conversations locally on a GPU
- control the voice of a spoken AI assistant with an audio prompt
- define an assistant's role and behavior via text prompts
- evaluate conversational speech models offline with wav files
- serve a low-latency spoken dialogue model over a web UI

## When to choose
- you need natural, low-latency spoken interaction with voice and role control
- you want a full-duplex model that can listen and speak simultaneously
- you have a CUDA GPU and want to self-host a conversational speech model
- you want to build on Moshi-compatible speech model weights

## When to avoid
- you only need text-based chat without audio
- you have no GPU and need high-throughput production inference
- you need a managed cloud speech API rather than self-hosted weights
- you need non-English conversation support without verifying language coverage

## Facets
- artifact type: library
- maturity: active
- function: speech-recognition, tts, llm-inference, machine-learning, http-server
- domain: speech-processing, artificial-intelligence, large-language-models, chatbots
- platform: python
- tags: full-duplex, speech-to-speech, voice-cloning, persona-control, real-time, moshi, linux, gpu, web-server, docker

## Member repositories
- NVIDIA/personaplex (main) score 47

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:10:41.901938+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 2026-08-29T17:19:04.729686+00:00, confidence not recorded.
  - readme: https://github.com/NVIDIA/personaplex (fetched 2026-08-28T04:10:41.901938+00:00, sha ce299a48c55a)
- Data as of 2026-08-30T08:39:29.467469+00:00.