# lipku/LiveTalking

Real time interactive streaming digital human

Repository: https://github.com/lipku/LiveTalking
Canonical: https://ross.abutalabs.com/products/livetalking
Homepage: https://www.livetalking.ai
Language: Python
License: Apache-2.0
License Family: permissive
Topics: digital-human, nerf, realtime, streaming, aigc, digihuman, musetalk, virtualhumans, wav2lip, er-nerf, lip-sync, talking-head, metahuman-stream
Last push: 2026-08-20T08:28:42+00:00

## Health v2 (maintenance only)
Score: 89/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 98, release rhythm 89, longevity 70
- inputs: {"age_days": 989, "days_push": 13, "days_rel": 75, "gap_med": 14, "n_releases_24m": 6}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 9238, forks 1459 (observed 2026-08-28T04:10:29.006396+00:00)

## What it is
LiveTalking is an open-source real-time interactive streaming digital human engine that drives a talking-head avatar from text or audio with lip-sync, optionally combined with LLM and TTS for conversational avatars. It supports multiple models (Wav2Lip, MuseTalk, ER-NeRF, Ultralight-Digital-Human), voice cloning, interruption, and outputs via WebRTC, RTMP, or virtual camera.

## Use cases
- build a real-time talking avatar that lip-syncs to speech
- create a 24/7 AI virtual streamer for live commerce
- deploy an AI digital human customer service agent
- generate talking-head videos from scripts via API
- build an interactive virtual presenter for education or kiosks
- integrate an LLM-powered avatar with voice conversation and interruption

## When to choose
- you need a self-hosted, commercially usable (Apache-2.0) real-time digital human with low latency
- you want to switch between multiple lip-sync models like Wav2Lip, MuseTalk, or ER-NeRF
- you need streaming output over WebRTC/RTMP or a virtual camera
- you want voice cloning, interruption support, and HTTP API integration with LLM/TTS pipelines

## When to avoid
- you only need offline video generation of talking heads without real-time streaming
- you lack a CUDA-capable GPU, since inference requires GPU acceleration
- you need enterprise features like transparent backgrounds, multi-avatar scenes, or wake-word interruption, which are in the paid commercial edition
- you want a no-code hosted SaaS rather than a deployable engine

## Facets
- artifact type: framework
- maturity: active
- function: tts, speech-recognition, llm-inference, video-processing, audio-processing, machine-learning, streaming, api-framework, chatbot
- domain: artificial-intelligence, computer-vision, deep-learning, web-development, media, self-hosted
- platform: windows, python, self-hosted
- tags: digital-human, talking-head, lip-sync, wav2lip, musetalk, nerf, webrtc, rtmp, voice-cloning, virtual-streamer, realtime-streaming, avatar, video, audio, linux, macos, docker, gpu, web-server

## Member repositories
- lipku/LiveTalking (main) score 89

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:10:29.006396+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 2026-08-29T17:23:15.124345+00:00, confidence not recorded.
  - readme: https://github.com/lipku/LiveTalking (fetched 2026-08-28T04:10:29.006396+00:00, sha abbeec171dab)
  - homepage: https://www.livetalking.ai (fetched 2026-08-29T08:23:02.923683+00:00, sha 20d5af8571dd)
  - site_page: https://doc.livetalking.ai/docs/service (fetched 2026-08-29T08:23:02.932852+00:00, sha 8d18591f9834)
  - site_page: https://doc.livetalking.ai (fetched 2026-08-29T08:23:02.934594+00:00, sha 22d7bf84944a)
- Data as of 2026-08-30T08:39:29.467469+00:00.