Weaving cosmic threads...
Weaving cosmic threads...
Voice-first creation with Guardian personas. Record a thought, hear it back in character. One-step voice-to-AI — no typing required.
Select a Guardian · hear the voice
Choose who reads your AI responses. Each persona has a distinct voice that matches their character.
From quick notes to studio voiceovers. Thinking, publishing, and workflow — voice adapts to your intent.
Quick thought capture in 60 seconds
5-minute session with action extraction
Speak a prompt, dispatch an agent
Broadcast-quality 48kHz recording
Studio production 48kHz/24bit
Speak an issue, create a ticket
Describe changes, get a commit
Daily standup via voice log
Not just recording — intelligent voice infrastructure for the AI age.
Auto-selects your best microphone. Broadcast > Studio > Gaming > Headset > Laptop.
Groq Whisper (free, <1s) with local Whisper fallback. Never pay for transcription.
6 Guardian personas with distinct voices. Lumina guides, Draconia commands, Lyria whispers.
Real-time filler detection, sentence analysis, vocabulary richness scoring.
Speak and send in one action. No typing, no editing — pure voice-to-AI flow.
Pure Node.js + ffmpeg. No bloated packages. Works on Windows, Mac, Linux.
Web, CLI, Claude Code, OpenCode, Cursor — the same voice intelligence in every environment. Zero npm dependencies. Cross-platform.
$ npx @arcanea/voice
Voice Note | 1m | 16kHz | BROADCAST | groq
Ctrl+C to stop | Shure MV6
Captured 847KB
Transcribing... [groq]
[COACH] 142w 8s 0f
Copied to clipboard
Five sovereign conductors. Luminor-mapped. Core public integrations only. All attest via SIP.
Stop typing prompts. Start speaking them.
Open Voice Chat