← Back to Skills Marketplace
wuxxin

Local Whisper (cpp)

by wuxxin · GitHub ↗ · v1.0.0
cross-platform ✓ Security Clean
1775
Downloads
2
Stars
10
Active Installs
1
Versions
Install in OpenClaw
/install local-whisper-cpp
Description
Local speech-to-text using whisper-cli (whisper.cpp).
README (SKILL.md)

Local Whisper (cpp)

Transcribe audio files locally using whisper-cli and the large-v3-turbo model.

Usage

You can use the wrapper script:

  • scripts/whisper-local.sh \x3Caudio-file>

Or call the binary directly:

  • whisper-cli -m /usr/share/whisper.cpp-model-large-v3-turbo/ggml-large-v3-turbo.bin -f \x3Cfile> -l auto -nt

Scripts

  • Location: scripts/whisper-local.sh (inside skill folder)
  • Model: /usr/share/whisper.cpp-model-large-v3-turbo/ggml-large-v3-turbo.bin
  • GPU: Enabled via whisper-cli.

Setup

Download the model to /usr/share/whisper.cpp-model-large-v3-turbo/:

wget https://huggingface.co/ggerganov/whisper.cpp/resolve/main/ggml-large-v3-turbo.bin?download=true -O /usr/share/whisper.cpp-model-large-v3-turbo/ggml-large-v3-turbo.bin
Usage Guidance
This skill appears to be what it says: a small wrapper around a local whisper-cli. Before installing/using it, ensure you have a trusted whisper-cli binary installed from an official source, and download the model from Hugging Face (as documented). Note that placing the model under /usr/share may require sudo or root access and significant disk space; you can instead store the model in a user-writable path and update the script/command to point to that location. Verify whisper-cli supports the flags used (-fa, -nt, GPU options) on your system. If you are uncomfortable granting write access to /usr/share, modify the instructions to use a safe directory.
Capability Analysis
Type: OpenClaw Skill Name: local-whisper-cpp Version: 1.0.0 The skill bundle is benign. It provides a wrapper script (`scripts/whisper-local.sh`) to run the `whisper-cli` tool for local speech-to-text transcription. The `SKILL.md` documentation includes a `wget` command to download a large language model from `huggingface.co`, which is a legitimate source for ML models and a standard setup step for such applications. There is no evidence of prompt injection, data exfiltration, malicious execution, or persistence mechanisms.
Capability Assessment
Purpose & Capability
Name/description match the implementation: the skill runs whisper-cli against a local model file. The only declared requirement is the whisper-cli binary, which is exactly what is needed for local transcription.
Instruction Scope
SKILL.md instructs downloading the model from a Hugging Face URL and storing it at /usr/share/whisper.cpp-model-large-v3-turbo/. That path may require elevated privileges (sudo) to write; the skill does not declare or request any config paths or elevated privileges explicitly. Runtime instructions otherwise only invoke whisper-cli on a supplied audio file and do not access unrelated files or env vars.
Install Mechanism
There is no automatic install spec (instruction-only). The provided model download uses a Hugging Face URL (a known host). No untrusted shorteners or personal servers are used, and the script itself is a small local wrapper.
Credentials
No environment variables, secrets, or unrelated credentials are requested. The model path is hard-coded under /usr/share but no credentials are required to run the skill.
Persistence & Privilege
always is false and the skill does not request persistent elevated privileges or modify other skills/config. It is user-invocable and can be run on demand.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install local-whisper-cpp
  3. After installation, invoke the skill by name or use /local-whisper-cpp
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial release
Metadata
Slug local-whisper-cpp
Version 1.0.0
License
All-time Installs 12
Active Installs 10
Total Versions 1
Frequently Asked Questions

What is Local Whisper (cpp)?

Local speech-to-text using whisper-cli (whisper.cpp). It is an AI Agent Skill for Claude Code / OpenClaw, with 1775 downloads so far.

How do I install Local Whisper (cpp)?

Run "/install local-whisper-cpp" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Local Whisper (cpp) free?

Yes, Local Whisper (cpp) is completely free (open-source). You can download, install and use it at no cost.

Which platforms does Local Whisper (cpp) support?

Local Whisper (cpp) is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Local Whisper (cpp)?

It is built and maintained by wuxxin (@wuxxin); the current version is v1.0.0.

💬 Comments