use-local-whisper

Switch WhatsApp voice transcription from OpenAI Whisper API to local whisper.cpp.

3|Updated Apr 19, 2026
One-click install
npx skills add https://github.com/OfirGavish/NinjaClaw-Nano --skill use-local-whisper-ofirgavish
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: use-local-whisper
Source: https://github.com/OfirGavish/NinjaClaw-Nano/tree/main/.copilot/skills/use-local-whisper
Command: npx skills add https://github.com/OfirGavish/NinjaClaw-Nano --skill use-local-whisper-ofirgavish

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Switches voice transcription from OpenAI Whisper API to local whisper.cpp, enabling on-device transcription with no API keys or network access. WhatsApp only for now; the transcription module uses Baileys types for audio download and requires the voice-transcription skill to be applied first.

Core Features & Use Cases

  • Local transcription: Transcribes audio on-device using whisper.cpp, removing cloud dependencies and data exposure.
  • Platform support: Currently supports WhatsApp; other channels require channel-specific audio download logic.
  • Prerequisite chaining: Requires the voice-transcription skill to be applied first to ensure proper channel integration.

Quick Start

Apply this skill to switch transcription to local whisper and test by sending a voice note in WhatsApp.

Frequently Asked Questions about use-local-whisper

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run offline voice transcription for WhatsApp on macOS Apple Silicon?

You can run offline voice transcription on macOS Apple Silicon by switching from the OpenAI Whisper API to local whisper.cpp, invoking whisper-cli to transcribe WhatsApp audio directly on-device.

Can I use whisper.cpp for on-device transcription without an OpenAI API key?

Yes, using whisper.cpp for on-device transcription removes cloud dependencies entirely, allowing you to transcribe WhatsApp voice notes locally without needing an OpenAI API key or network access.

What prerequisites are needed to set up local whisper transcription?

To set up local whisper transcription, you must first apply the voice-transcription skill for WhatsApp channel integration, install whisper-cpp, and ensure a GGML model exists at data/models.

Does local whisper.cpp voice transcription work with channels other than WhatsApp?

Local whisper.cpp voice transcription currently supports WhatsApp only; other channels require channel-specific audio download logic because the transcription module uses Baileys types for audio retrieval.

How do I verify that local whisper transcription is working correctly?

You can verify local whisper transcription by applying the skill configuration to invoke whisper-cli, ensuring the GGML model is downloaded, and testing the setup by sending a voice note in WhatsApp.

Why does offline transcription require the voice-transcription skill to be applied first?

Offline transcription requires the voice-transcription skill first to establish proper WhatsApp channel integration and audio download mechanisms before switching the backend to local whisper.cpp.