Troubleshooting

Troubleshooting

Voice

"whisper-cli not found" whisper.cpp isn't installed. Run brew install whisper-cpp and restart Tome — or switch to Apple on-device in Preferences → Voice (macOS).

"Speech model not downloaded" The whisper model file is missing. Click Download speech model in Preferences → Voice, or fetch it manually with the command Tome shows.

Microphone unavailable / permission denied Grant microphone access to Tome in System Settings → Privacy & Security → Microphone, then retry.

On-device recognition unavailable Your Mac doesn't support Apple's on-device dictation for the current language. Switch the engine to whisper.cpp in Preferences → Voice.

Transcription timed out The local model took too long (large model, slow machine). The warm-up preference can help; or switch engines.

Assistant

"…needs — export it in your shell and restart" The provider key isn't in your login shell. Export it, restart Tome, or pick a different provider in Preferences → Assistant.

"Chat credentials rejected" The key is set but rejected. Check it's current and valid for the provider.

Containment

A host is blocked The pane is gapped and that host isn't on the allowlist. Unlock the pane (if you intend it) or add the host via the egress UI.

Unlock didn't stick / relocked Relock timers may have fired. Check the security settings in Preferences.

Layout

Layout didn't restore Fixed in v0.3.0. If you're on a newer build and it recurs, file an issue with your workspace's layout JSON.

Where app data lives

Tome stores its state under the OS app-data directory (config store, downloaded models, egress configs). On macOS this is under ~/Library/Application Support/tome/.