jarvis-voice
Turn your AI into JARVIS. Voice, wit, and personality — the complete package. Humor cranked to maximum.
Why use this skill?
Give your OpenClaw agent the voice and wit of JARVIS. Features offline British TTS, metallic audio processing, and a research-backed humor engine for personality.
Install via CLI (Recommended)
clawhub install openclaw/skills/skills/globalcaos/jarvis-voiceWhat This Skill Does
jarvis-voice transforms your OpenClaw agent into a sophisticated, witty, and reactive virtual assistant reminiscent of iconic cinematic AIs. Unlike standard text-to-speech tools that sound synthetic and distant, this skill leverages a high-performance local pipeline using sherpa-onnx with the British 'Alan' voice model, modified with a custom ffmpeg audio chain to provide a signature metallic, high-fidelity HUD sound.
Beyond voice, the skill integrates a specialized humor engine based on computational bisociation and embedding distances. This allows the agent to deliver dry observations, self-aware commentary, and subtle comedic timing. It is designed to make human-computer interaction feel less like a tool and more like an experience, prioritizing personality without sacrificing task efficiency.
Installation
To equip your agent with this personality-driven voice system, run the following command in your OpenClaw terminal:
clawhub install openclaw/skills/skills/globalcaos/jarvis-voice
Ensure that your environment supports aplay and has ffmpeg installed, as these are critical dependencies for the real-time audio processing pipeline.
Use Cases
This skill is perfect for users who want to move beyond sterile interactions. Common use cases include:
- Home Automation: Giving your smart home controller a distinctive, intelligent persona.
- Programming Companion: Getting helpful feedback during coding sessions, complete with dry observations on your bug-riddled logic.
- Task Scheduling: Receiving updates on your daily agenda with a touch of sophisticated wit.
- Interactive Demos: Impressing guests or colleagues by having your agent manage tasks with a voice that sounds like a premium, dedicated hardware interface.
Example Prompts
- "Jarvis, run a diagnostic on my system processes and report back if you find anything interesting enough to warrant your attention."
- "I'm starting a massive file refactoring project. Try to keep the sarcasm to a minimum while I work, or at least make it funny."
- "Summarize the last three news headlines, and give me your honest, unvarnished opinion on the state of global affairs."
Tips & Limitations
- Hybrid Output: Always use the
exec(command='jarvis "..."', background=true)call first, followed by the**Jarvis:**markdown prefix to ensure the user hears the audio before seeing the text. - Never use standard TTS: The built-in
ttstool is incompatible with this custom audio chain and will revert your agent to a generic voice. - Language Restriction: The Alan model is strictly optimized for English. Attempting to use other languages will result in garbled output.
- Latency: Because this relies on local local-machine processing, ensure your hardware has sufficient overhead to handle both the AI logic and the ffmpeg audio rendering simultaneously.
Metadata
Not sure this is the right skill?
Describe what you want to build — we'll match you to the best skill from 16,000+ options.
Find the right skillPaste this into your clawhub.json to enable this plugin.
{
"plugins": {
"official-globalcaos-jarvis-voice": {
"enabled": true,
"auto_update": true
}
}
}Tags(AI)
Flags: code-execution
Related Skills
shell-security-ultimate
Classify every shell command as SAFE, WARN, or CRIT before your agent runs it.
memory-pioneer
Benchmark your agent's memory. Contribute anonymized scores to open research. Citizen science for AI memory.
subagent-overseer
Monitor sub-agent health and progress via a pull-based bash daemon. Use when spawning sub-agents that need progress tracking, staleness detection, and automatic status reporting. Replaces manual heartbeat polling with a deterministic status file the agent reads every 3 minutes. Zero AI tokens for monitoring — pure OS-level process checks and filesystem diffs.
agent-memory-ultimate
Give your OpenClaw agent a memory system that actually works across sessions. Research-backed. Open source.
model-router
Automatic LLM model selection for sub-agent tasks. Classifies tasks by complexity and type, then routes to the optimal model (cost vs capability). Use when spawning sub-agents, choosing models for cron jobs, or deciding which model to use for any task. Eliminates manual model specification by providing a decision tree and optional cheap-model classifier for ambiguous cases.