实时语音 AI agent——语音进语音出的对话系统,接 Whisper、ElevenLabs、OpenAI Realtime。
这 30 个语音智能体里，首选 qwen-audio-agent（安全扫描评级 SAFE，★ 3.0k）。 A realtime voice runtime that keeps Agents talking, working, and present. Real-time Voice Runtime for AI Agents
语音智能体工具是一类专注于帮助开发者和团队解决 语音智能体 相关任务的 AI 驱动软件。这些工具通常以开源形式发布在 GitHub 上，支持通过 MCP（Model Context Protocol）、Claude Skills 或独立 Agent 框架集成到现有工作流中。在 Agent Skills Hub 上，我们收录了 30 个经过质量评分的 语音智能体工具，覆盖TypeScript、Rust、JavaScript等主流编程语言。
在 2026 年，AI Agent 生态系统正在快速成熟。语音智能体工具能够显著提升开发效率：自动化重复任务、减少人为错误、并提供智能建议。排名前三的工具——openlive、agents-kit、cli——平均获得了 4,676 个 GitHub Star，体现了开发者社区的高度认可。其中 25 个工具提供了明确的开源许可证，确保你可以自由使用和修改。
选择 语音智能体工具时，建议考虑以下因素：1️⃣ 社区活跃度（GitHub Star 数和最近提交频率）；2️⃣ 集成方式（是否支持 MCP、Claude 或你使用的 Agent 框架）；3️⃣ 编程语言兼容性（本列表中最常见的语言是 TypeScript）；4️⃣ 质量评分（Agent Skills Hub 的综合评分考量了代码质量、文档完整性和维护活跃度）。我们的推荐是从 openlive 开始——它在 Star 数和质量评分上都名列前茅。
Opensource, on-device voice + vision layer for AI agents. Bring any model or coding agent; the whole speech loop (VAD, STT, TTS, barge-in) runs locally. An open alternative to ElevenLabs Agents, Gemini Live, and OpenAI Realtime.
React components for voice agents, chat, and generative UI. LiveKit, ElevenLabs, and OrbKit in one kit. Copy the source and make it yours.
npx shadcn@latest add https://agents-ui.github.io/agents-kit/c/agent-generative-surface.json
Command Line Interface to interact with ElevenLabs voice agents
npm install -g @elevenlabs/cli
A realtime voice runtime that keeps Agents talking, working, and present. Real-time Voice Runtime for AI Agents
npm install -g qwen-audio-agent
Multi-modal voice agent for your AI agents. Talk with Claude Code, Codex, and others by voice: answer their prompts, give instructions, ask what they did. Or just nod.
git clone https://github.com/spaceamoeba-t/tapq.git
用于使用 LiveKit 构建、测试和发布语音 agent 的 Agent Skills 与 Claude Code 插件
npx skills add livekit/agent-skills
Open-source AI phone API for voice agents. Give Claude, Cursor, OpenClaw, or any LLM a real phone number — outbound/inbound calls, SMS, transcripts, MCP, DTMF/IVR. FastAPI + SignalWire + Deepgram. Self-hostable alternative to Twilio, Vapi, and Retell.
git clone https://github.com/agentlineHQ/AgentLine.git
JARVIS — a voice assistant for Claude Code. Talk to your Mac and he brainstorms a project with you, builds it, and tells you out loud when a Claude Code session needs you. macOS, on your Claude subscription, no API key.
Rust Agent Development Kit (ADK-Rust): Build AI agents in Rust with modular components for models, tools, memory, realtime voice, and more. ADK-Rust is a flexible framework for developing AI agents with simplicity and power. Model-agnostic, deployment-agnostic, optimized for frontier AI models. Includes support for real-time voice agents.
Open-source local AI audio & music studio for macOS & Windows. TTS, voice cloning, songs with lyrics, and multi-track scores — 100% offline, zero API fees.
Infrastructure for the next generation of voice agents, designed to provide universal memory. It is divided into a left brain and a right brain, storing information and emotions respectively, while a fully streaming architecture eliminates latency at the fundamental level.
git clone https://github.com/lang-jiaqi/Voicemem_open.git
💖🧸 Self hosted, you-owned Grok Companion, a container of souls of waifu, cyber livings to bring them into our worlds, wishing to achieve Neuro-sama's altitude. Capable of realtime voice chat, Minecraft, Factorio playing. Web / macOS / Windows supported.
TypeScript AI SDK and agent framework with zero dependencies. Build production AI agents and multi-agent swarms: durable execution, long-term memory, hybrid RAG, MCP client/server, A2A, human-in-the-loop approvals, CodeAct sandboxes, realtime voice. One streaming API, 29 providers: Claude, GPT, Gemini, Grok, DeepSeek, Ollama. Node, Bun, Deno, edge.
npm install @deuz-sdk/core # the runtime
Voice-driven multi-agent assistant — Realtime API + CLI runners + multi-channel orchestration.
git clone https://github.com/OpenYabby/OpenYabby.git
Open-source MCP server for safe, stateful AI-agent phone calls with OpenAI Realtime and Twilio.
The One Dashboard every Voice Agent needs
pip install "voicegateway[dashboard]"
Fast, free voice assistant for your Mac. No API key: runs on your Claude Pro or Max plan. Sends background agents for big jobs, remembers you and learns your shortcuts. Hold ⌃⌥Space and talk.
Native Android AI agent: use your own API keys (OpenAI/Anthropic/Gemini), opencode-style modes, device tools, Shizuku shell, screen analysis, web search, custom Skills, voice assistant
Nexus | Take over your Android voice assistant with a custom AI agent — connect any LLM (Claude, GPT, Gemini), use MCP tools, Shell/SSH, and Root/Shizuku automation
open sourced personal agent + dual-brain voice assistant. 开源个人助手 + 前后脑语音助手. restrained proactivity, wide context. 主动但不啰嗦，懂你所想，帮你所需
npm install --global nova-audio-agent@0.2.3
open sourced personal agent + dual-brain voice assistant. 开源个人助手 + 前后脑语音助手. restrained proactivity, wide context. 主动但不啰嗦，懂你所想，帮你所需
npm install --global nova-personal-agent@latest
Open-source voice-AI SDK. The Vapi/Retell alternative for builders who want to own the stack. Give your AI agent a phone number in 4 lines — Python and TypeScript, MIT licensed, Twilio, Telnyx, and Plivo.
npx skills add patterai/skills
开源的 agent 视频制作系统：12 条制作流水线、100+ 工具、700+ skill 与知识文件
git clone https://github.com/calesthio/OpenMontage.git
Voice-powered AI assistant platform — connect any LLM, any TTS, with a live web canvas, music generation, and agent orchestration using openclaw. Install: npx openvoiceui setup
npx openvoiceui setup # interactive wizard — walks you through API keys + builds Docker images
Hermes Talk: realtime duplex voice for Hermes Agent. Rides your ChatGPT or X Premium subscription, no API key needed. OpenAI, Grok, Gemini. Terminal, Discord, dashboard.
pip install "hermes-talk[audio]" # mic + speaker support (sounddevice); skip if dashboard-only
A lightweight, powerful framework for multi-agent workflows and voice agents
面向 Claude Code 的 AI 原生视频制作工具包
git clone https://github.com/digitalsamba/claude-code-video-toolkit.git
The official ElevenLabs MCP server
git clone https://github.com/elevenlabs/elevenlabs-mcp
Like the macOS say command, but with a modern voice.
brew install steipete/tap/sag # auto-taps steipete/tap
| 工具 | 星标 | 评分 |
|---|---|---|
| openlive | ★ 329 | 66 |
| agents-kit | ★ 118 | 57 |
| cli | ★ 102 | 63 |
| qwen-audio-agent | ★ 3.0k | 77 |
| tapq | ★ 217 | 60 |
| agent-skills | ★ 71 | 72 |
| AgentLine | ★ 56 | 74 |
| bolna | ★ 800 | 67 |
| jarvis | ★ 736 | 70 |
| adk-rust | ★ 719 | 68 |
| yovoice | ★ 1.8k | 68 |
| VoiceMem | ★ 1.6k | 69 |
| airi | ★ 50.6k | 75 |
| Deuz-SDK | ★ 570 | 62 |
| OpenYabby | ★ 125 | 47 |
| agent-call | ★ 97 | 57 |
| voicegateway | ★ 54 | 58 |
| atticus-mac | ★ 222 | 71 |
| BetterAIChat | ★ 218 | 70 |
| com.niki914.nexus.agentic | ★ 155 | 61 |
2026 年顶尖的语音智能体工具是 openlive, agents-kit, cli。Agent Skills Hub 按 GitHub Star、质量分（含完整度、示例、agent 就绪度等 6 个维度）和近期活跃度对 30 个选项排序，每 8 小时基于实时 GitHub 数据重建。
openlive（329 stars）是通用语音智能体工作流中采用最广的选择，用 TypeScript 编写。agents-kit（118 stars）是有力的替代。按你现有技术栈来选：匹配团队已用的语言和运行时，把集成成本降到最低。拿不准就从 openlive 开始——它社区最深、线上示例最多。
遇到以下情况建议别用现成的语音智能体工具：（1）需求需要工具插件系统不支持的深度定制；（2）有严格合规要求、禁止第三方依赖；（3）工具维护停滞（上次提交超过 6 个月）；（4）数据量小到写个 50 行脚本比学这工具还省。对每天 100 次请求以上的生产工作流，维护良好的工具省下的时间通常盖过定制损失。
语音智能体专门聚焦语音智能体相关的具体任务；文本转语音与配音是相关但不同的类别——那类工具见 https://agentskillshub.top/best/text-to-speech/。两者常出现在同一条 agent 流水线里，却解决不同问题：主目标是具体任务时选语音智能体，工作流更宽泛时选文本转语音与配音。
对多数团队是的。openlive 有 329 个 star 的社区验证，能处理你没想到的边界情况，还自带文档。只有当（1）需求高度非标准、（2）有安全/合规理由要避开开源依赖、或（3）维护负担小到（代码 <200 行）长期更省时，才值得自己造。盈亏平衡点通常在省下约 2-3 周开发时间。
列表里多数语音智能体工具是开源的，采用宽松许可（MIT、Apache 2.0）。少数在免费自托管内核之上提供付费托管/云版本。商用前务必查看每个工具 GitHub 仓库的 LICENSE——有些用 AGPL 或非商业限制，可能不适合你的部署模式。
评级来自基于 SlowMist agent 安全分类法的规则扫描：42 条公开规则，覆盖凭证窃取、数据外泄、curl | sh 安装等。它是第一层扫描，不是人工审计 — 我们如实说明，不夸大。
问题规模有独立研究佐证：Liu et al. (2026) 对 31,132 个 agent skill 的研究发现，26.1% 含有安全漏洞。我们自己的全目录普查已作为可引用的开放数据集发布。
信源与负责人：