Sag

**Quick judgement**: `sag` is a focused, well-scoped CLI wrapper around ElevenLabs TTS with a mac-style `say` UX. It’s a good fit if you already use ElevenLabs and want a simple way for agents or scripts to generate local audio replies. It does not provide its own TTS; it relies entirely on the ElevenLabs API and your API key. **Recommended scenarios** - Agent setups where you occasionally want to respond “in voice” (e.g., generate an MP3 and attach it to a message). - Developers already using ElevenLabs who want a fast, terminal-native tool for testing voices and prompts. - Local tooling or automation where text needs to be spoken aloud or turned into audio files on demand. **Key risks and limitations** - **External dependency**: All TTS runs through ElevenLabs; outages, latency, and costs are governed by that service. - **Privacy/compliance**: Any text you convert to speech is sent to ElevenLabs; unsuitable for highly sensitive data without proper review. - **Platform assumptions**: UX is modeled on macOS `say`; while it should work as a CLI tool, exact behavior and playback may vary by environment. - **Abuse potential**: As with any high-quality TTS, there is potential for misuse (e.g., impersonation) if paired with certain voices; governance must be handled at the deployment level, since this tool is a thin wrapper around the API.