| name | cli-anything-copyspeak |
| description | Operate CopySpeak desktop TTS from a stateful CLI: manage projects, queue text, invoke real TTS backends, and verify audio output. |
cli-anything-copyspeak
Use this skill when an agent needs to drive CopySpeak without the GUI.
Install
cd /d/GitHub/copyspeak-tts/agent-harness
pip install -e .
Hard dependencies: real CopySpeak (COPYSPEAK_EXE or built Tauri executable) and a real TTS backend (kittentts-cli.py, Piper, Kokoro, etc.). Exports fail loudly if the backend is missing.
Command pattern
Global flags come before command groups:
cli-anything-copyspeak --project project.json --json <group> <command>
Use --json for machine-readable output. Mutations auto-save; add --dry-run to suppress saving.
Commands
project new -o FILE [--name NAME] — create project JSON.project info — inspect current project.project set-config --engine ENGINE --voice VOICE --speed N --pitch N --volume N — update synthesis settings.queue add -t TEXT [--label LABEL] — enqueue text.queue list — list queued text.queue remove ITEM_ID — remove queued item.queue clear — clear queue.export text -t TEXT -o AUDIO [--overwrite] — synthesize one item using the real backend.export queue -o DIR [--overwrite] — synthesize all queued items.backend check — locate CopySpeak executable.backend launch — launch CopySpeak.
Examples
cli-anything-copyspeak project new -o demo.json --name Demo
cli-anything-copyspeak --project demo.json queue add -t "Read this aloud" --label intro
cli-anything-copyspeak --project demo.json --json queue list
cli-anything-copyspeak --project demo.json export queue -o audio --overwrite
The export path validates non-empty audio and basic magic bytes. Do not treat success as only subprocess exit; inspect returned output and file_size.
