#2222nd of 81 in VoiceFor voice, speech and music
xiaozhi-esp32-server
Backend server for xiaozhi-esp32 voice devices with ASR, LLM and TTS pipeline
- Stars
- 11.3k
- License
- MIT
- Last commit
- Nov 2026
- Last release
- Oct 2026
- Language
- JavaScript
9,246 on 21 Apr 202611,254 on 10 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
Self-hosted backend for the open-source xiaozhi-esp32 hardware, implemented in Python, Java and Vue per the Xiaozhi communication protocol. It connects devices over WebSocket or MQTT+UDP to configurable ASR, LLM, VLLM and TTS providers, with VAD, voiceprint speaker ID, MCP tool calling and RAGFlow knowledge bases. A minimal mode keeps config in files; the full mode adds a web console with multi-user and multi-agent management.
Who it is for: Owners of xiaozhi-esp32 devices who want their own voice backend
Strengths
- Many swappable providers for ASR, TTS, LLM and VLLM, plus any OpenAI-compatible endpoint
- Streaming ASR/TTS configuration; README claims about 2.5 s faster responses since 0.5.2
- Minimal install stores config in files and needs no database
- Supports MCP (client, server, access point), plugins, mem0ai/PowerMem memory and RAGFlow
Weaknesses
- Needs ESP32 hardware running xiaozhi-esp32 firmware to be useful
- README says features are incomplete and not security-tested; not for production
- Voiceprint needs a separate voiceprint-api service plus MySQL
- README is primarily Chinese; local FunASR is memory-heavy
What it needs
- GPU optional
- Needs voiceprint-api (optional), MySQL (voiceprint), database (full-module install), RAGFlow (optional), third-party ASR/LLM/TTS APIs
- Models: FunASR, SherpaASR, VoskASR, SileroVAD, EdgeTTS
Quick answers
Is xiaozhi-esp32-server open source?
Yes. Its code is on GitHub under the MIT license.
Does xiaozhi-esp32-server need a GPU?
No. It runs on a CPU and can use a GPU when you have one.
Also in Voice
See all 81| Rank | Project | Score |
|---|---|---|
| 1 | voicestudioDesktop app for voice cloning, dubbing, dictation and audiobooks | 98 out of 100 |
| 2 | handyOffline push-to-talk dictation app that pastes transcripts into any text field | 91 out of 100 |
| 3 | whisper.cppC/C++ speech-to-text engine running OpenAI Whisper models locally | 90 out of 100 |
| #4 | xiaozhi-esp32ESP32 firmware for a voice chatbot with MCP device and cloud tool control | 83 out of 100 |
| #5 | buzzOffline desktop app for transcribing and translating audio with Whisper | 83 out of 100 |