voice-agents
Voice agents represent the frontier of AI interaction - humans speaking naturally with AI systems.
npx skills add majiayu000/claude-skill-registry --skill voice-agents-sickn33-antigravity-awesome-3 --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
What it does
Conveys capabilities for voice-enabled agents, including speech-to-speech and pipeline architectures, aiming for end-to-end latency under 800ms and smooth turn-taking with interruption handling. It lists procedures and configurations for latency targets, VAD types, and example implementations across OpenAI Realtime API and various STT/TTS tools. The goal is real-time, natural conversations via voice interfaces with structured guidance on components and patterns.
How it works
It outlines two architectures:
- Speech-to-Speech: direct audio-to-audio flow with modules for input audio, S2S model, and agent audio, emphasizing sub-500ms latency and preservation of emotion.
- Pipeline: separate STT → LLM → TTS steps for full control and easier debugging, acknowledging higher typical latency (700-1200ms). It provides concrete tool options for each stage, including OpenAI Realtime API and Pipecat for orchestration, Deepgram/Nova-3 and Whisper for STT, ElevenLabs/OpenAI/Deepgram TTS, and frameworks like Pipecat, Vapi, Retell AI. It also describes latency targets, TTFT/TTFA timings, and a breakdown of contributing components. It includes example code blocks for configuring a realtime API session, and production-style pipelines, plus patterns for optimization (streaming, pre-computation, edge deployment).
When to use it
Triggered by needs for low-latency, natural voice conversations with constraints on end-to-end latency (<800ms), minimal jitter, effective interruption handling, and realistic conversational flow. It specifies that latency is the primary constraint and emphasizes exact timing goals (e.g., TTFT < 300ms, end-to-end < 800ms) and adaptive VAD behavior.
What it can touch
Declared tools and components include:
- OpenAI Realtime API (for S2S and pipeline usage)
- Pipecat (orchestrator)
- STT: OpenAI Whisper, Deepgram Nova-3, AssemblyAI
- TTS: ElevenLabs, OpenAI TTS, Deepgram Aura-2
- Frameworks: Pipecat, Vapi, Retell AI
- VAD options: Silero VAD via Pipecat, OpenAI Semantic VAD These tools and integrations are described as the mechanisms by which the agent operates.
Caveats
Stated risks labeled as safe; emphasizes latency and jitter considerations as critical factors, with explicit latency targets and potential trade-offs (e.g., pipeline latency vs. control, streaming vs. full transcripts). It includes cautions about debugging and auditing in the pipeline and notes possible disadvantages (e.g., higher latency for the pipeline approach).
# Voice Agents Voice agents represent the frontier of AI interaction - humans speaking naturally with AI systems. The challenge isn't just speech recognition and synthesis, it's achieving natural conversation flow with sub-800ms latency while handling interruptions, background noise, and emotional nuance. This skill covers two architectures: speech-to-speech (OpenAI Realtime API, lowest latency, most natural) and pipeline (STT→LLM→TTS, more control, easier to debug). Key insight: latency is the constraint. Humans expect responses in 500ms. Every millisecond matters. 84% of organizations are increasing voice AI budgets in 2025. This is the year voice agents go mainstream. ## Principles - Latency is the constraint - target <800ms end-to-end - Jitter (variance) matters as much as absolute latency - VAD quality determines conversation flow - Interruption handling makes or breaks the experience - Start with focused MVP, iterate based on real conversations - Combine best-in-class components (Deepgram STT + ElevenLabs TTS) ## Capabilities - voice-agents - speech-to-speech - speech-to-text - text-to-speech - conversational-ai - voice-activity-detection - turn-taking - barge-in-detection -
- Principles
- Capabilities
- Scope
- Tooling
- Speechtospeech
- Speechtotext
- Texttospeech
- Frameworks
- Patterns
- Speech-to-Speech Architecture
- OpenAI Realtime API
- Use Cases:
- Pipeline Architecture
- Production Pipeline Example
What does the voice-agents skill do?
Voice agents represent the frontier of AI interaction - humans speaking naturally with AI systems.
How do I install it?
Run `npx skills add majiayu000/claude-skill-registry --skill voice-agents-sickn33-antigravity-awesome-3 --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From majiayu000/claude-skill-registry, a repository with 534 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.