v0.5.2

rohitg00/agentmemoryv0.5.2Sep 25, 2026by andimarafioti

Upgrading?

AI Summary

Improves the user experience for the GGML CLI by making native diagnostics quiet by default, keeping only the final success message visible while maintaining model download progress.

Key Highlights

  • GGML CLI commands are now quiet by default
  • Native diagnostics no longer flood successful runs
  • Verbose mode available via `--verbose` for troubleshooting

New Features

  • Quieter CLI output for GGML commands
  • Improved verbose diagnostics for troubleshooting

Full Release Notes

## What's changed

- GGML CLI commands are quiet by default: native qwentts.cpp and Metal diagnostics no longer flood successful runs. The final `Wrote ...` message remains visible.
- Model download progress remains visible, and native diagnostics are shown if a command fails.
- Use `faster-qwen3-tts --verbose ...` to show native diagnostics while troubleshooting. Place `--verbose` before the subcommand.

The package version and demo pins are now 0.5.2.

Full changes: https://github.com/andimarafioti/faster-qwen3-tts/pull/142

Related Repositories