← Back to Skills Marketplace
tobihagemann

yap

by tobihagemann · GitHub ↗ · v1.0.1
darwin ✓ Security Clean
642
Downloads
0
Stars
1
Active Installs
2
Versions
Install in OpenClaw
/install yap
Description
Fast on-device speech-to-text transcription on macOS 26+ using Apple Speech.framework, supporting multiple languages and output formats without model downloads.
Usage Guidance
This skill appears to do what it says: run the 'yap' CLI to transcribe local audio on macOS. Before installing, review the Homebrew formula and the project's GitHub repo (finnvoor/tools/yap) to confirm the source and any install-time actions. Be aware the installed binary will have access to any files you give it for transcription. Also note that although the SKILL.md emphasizes 'no model download', Apple Speech may rely on system speech models which your machine or Apple may download or manage; if you need to avoid network activity, verify how your macOS handles Speech.framework models. If you don't trust the third-party tap, consider building from source or obtaining the binary from an audited release.
Capability Analysis
Type: OpenClaw Skill Name: yap Version: 1.0.1 The skill bundle defines a macOS speech-to-text tool called 'yap'. The `SKILL.md` file provides clear installation instructions using `brew install finnvoor/tools/yap` and usage examples for local file transcription. There are no signs of prompt injection, data exfiltration, malicious execution, persistence mechanisms, or obfuscation. All instructions and metadata are consistent with a benign utility skill.
Capability Assessment
Purpose & Capability
Name/description (on-device Apple Speech.framework transcription) match the declared requirement (a single 'yap' binary) and the SKILL.md usage examples. The macOS darwin restriction is appropriate.
Instruction Scope
SKILL.md only tells the agent to run the 'yap transcribe' CLI against user-supplied audio/video paths and lists expected flags. It does not instruct the agent to read unrelated files, access unrelated env vars, or exfiltrate data to external endpoints.
Install Mechanism
Install uses a Homebrew formula (finnvoor/tools/yap). Homebrew installation is a standard mechanism, but this is a third-party tap/formula rather than an official core formula. Installing third-party brew formulas can execute install scripts and pull arbitrary code/binaries, so you should inspect the formula or the project's GitHub releases before installing.
Credentials
The skill requires no environment variables or credentials and does not request config paths. This is proportional for a local transcription CLI. Note: the binary will be run with access to any files/paths the agent is asked to transcribe.
Persistence & Privilege
The skill is not always-enabled and does not request persistent system-wide privileges. It does not claim to modify other skills or global agent settings.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install yap
  3. After installation, invoke the skill by name or use /yap
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.1
Add OS declaration (darwin) to metadata
v1.0.0
Initial release: on-device STT via Apple Speech.framework (macOS 26+)
Metadata
Slug yap
Version 1.0.1
License
All-time Installs 1
Active Installs 1
Total Versions 2
Frequently Asked Questions

What is yap?

Fast on-device speech-to-text transcription on macOS 26+ using Apple Speech.framework, supporting multiple languages and output formats without model downloads. It is an AI Agent Skill for Claude Code / OpenClaw, with 642 downloads so far.

How do I install yap?

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

Is yap free?

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

Which platforms does yap support?

yap is cross-platform and runs anywhere OpenClaw / Claude Code is available (darwin).

Who created yap?

It is built and maintained by tobihagemann (@tobihagemann); the current version is v1.0.1.

💬 Comments