Skip to content

troubleshooting.md

Latest commit

 

History

History
131 lines (96 loc) · 1.99 KB

troubleshooting.md

File metadata and controls

131 lines (96 loc) · 1.99 KB

Troubleshooting

Common Issues

"Python environment not set up"

Run setup:

speak setup

If that fails, try forcing a reinstall:

speak setup --force

"Server not running" or connection errors

Clean up stale socket and restart:

speak daemon kill
speak health

Audio still playing after Ctrl+C

Kill any lingering audio processes:

pkill afplay

Generation hangs or times out

For long documents, use auto-chunking:

speak long-document.md --auto-chunk --output output.wav

If it still times out, increase the timeout:

speak document.md --timeout 600  # 10 minutes

"sox not found" when using --auto-chunk or concat

Install sox:

brew install sox

Poor audio quality

Try the full-precision model:

speak "Hello" --model mlx-community/chatterbox-turbo-fp16

Or adjust temperature (lower = more consistent, higher = more expressive):

speak "Hello" --temp 0.3  # More consistent
speak "Hello" --temp 0.7  # More expressive

Out of memory errors

Use a smaller model:

speak "Hello" --model mlx-community/chatterbox-turbo-4bit

Or stop other memory-intensive applications.

Health Check

Run a comprehensive health check:

speak health

This checks:

  • Python environment
  • Required packages
  • Server status
  • Audio device availability
  • Disk space

Logs

View today's log:

cat ~/.chatter/logs/speak_$(date +%Y-%m-%d).log

Enable verbose output for debugging:

speak "Hello" --verbose

Or set debug log level in config:

# ~/.chatter/config.toml
log_level = "debug"

Resetting Everything

If all else fails, reset the environment:

# Stop daemon
speak daemon kill

# Remove environment
rm -rf ~/.chatter/env

# Reinstall
speak setup

This preserves your config file and voice presets.

Getting Help

# Show all commands
speak --help

# Show help for a specific command
speak setup --help
speak concat --help