Feature/headless transcribe cli (for use in OpenClaw)

[mane81]

Before Submitting This PR

• I have searched existing issues and pull requests (including closed ones) to ensure this isn't a duplicate
• I have read CONTRIBUTING.md

If this is a feature or change that was previously closed/rejected:

• I have explained in the description below why this should be reconsidered
• I have gathered community feedback (link to discussion below)

Human Written Description

I built this for my personal local automation setup with OpenClaw.
My main use case is receiving voice messages (Telegram/WhatsApp) and having OpenClaw transcribe them reliably via a deterministic CLI call, instead of UI/hotkey interaction.
This also helps avoid running multiple separate STT stacks in parallel and keeps my workflow simpler while still using Handy’s local model selection.

Related Issues/Discussions

Fixes #N/A
Discussion: N/A (happy to open one if required before merge)

Community Feedback

No dedicated discussion thread yet.
If preferred, I can open one first and collect feedback there before merge.

Testing

Tested locally on macOS (Apple Silicon):

• cargo check -q passes
• cargo run -- --transcribe-file /path/to/audio.wav --format text returns transcript to stdout
• cargo run -- --transcribe-file /path/to/audio.wav --format json returns JSON output
• Verified default behavior uses selected model from settings_store.json
• Verified --model-id override works

Screenshots/Videos (if applicable)

N/A (CLI-only change)

AI Assistance

• No AI was used in this PR
• AI was used (please describe below)

If AI was used:

• Tools used: OpenClaw assistant (Codex 5.3)
• How extensively: Medium — AI helped with implementation/refactoring and wording; behavior and tests were validated by me.

[AlexanderYastrebov]

Check out https://k2-fsa.github.io/sherpa/onnx/index.html which is more suitable for automation and has a lot of tools and features.
See examples from https://k2-fsa.github.io/sherpa/onnx/pretrained_models/offline-transducer/nemo-transducer-models.html?highlight=parakeet specifically.