Troubleshooting

Problem: "Failed to fetch LM Studio models"

Solution:

1. Verify LM Studio is running
2. Check URL is http://localhost:1234/v1
3. Load a model in LM Studio first
4. Click "Refresh" in Application Settings

Problem: Whisper transcription errors

Solution:

1. Check GPU drivers if using CUDA
2. Try smaller model (base instead of large)
3. Verify video has audio track
4. Check logs/ folder for details

Problem: Transcription takes too long

Solution:

• Enable GPU acceleration in Transcribe settings (requires NVIDIA GPU)
• Use a smaller model (medium instead of large-v3)
• Close other GPU-intensive applications
• Update your GPU drivers

Problem: Metadata generation fails or times out

Solution by provider:

• Ollama: Make sure Ollama is running (ollama serve)
• OpenAI: Check your API key and account balance
• LM Studio: Ensure the local server is started
• Grimnir: Verify a model is downloaded and selected

Problem: Settings not persisting

Solution:

Auto-save is automatic! Changes save every 500ms. If issues persist:

1. Check file permissions in VideoProfiles/ folder
2. Verify app_settings.json is writable
3. Check logs/ for errors

Problem: YouTube upload errors

Solution:

1. Verify OAuth credentials are configured (see YouTube OAuth Setup)
2. Re-authenticate if token expired
3. Check internet connection
4. Verify video meets YouTube requirements
5. Check YouTube's daily upload limits

Problem: Custom thumbnail doesn't appear after upload

Solution:

• YouTube can take up to 24 hours to process custom thumbnails
• If it persists, try re-uploading the thumbnail manually through YouTube Studio
• Verify your account is eligible for custom thumbnails (requires phone verification)

Problem: AI descriptions talk about game sounds instead of your commentary

Solution:

• Check Primary radio button is selected for YOUR track
• Verify track name includes "Mic", "Host", or "Commentary"
• Re-transcribe if you changed track names

See How It Works for track naming best practices.

Problem: SRT captions have in-game character voices

Solution:

• Ensure game track is named "System Audio" or "Game Audio"
• Track 1 (index 0) should ALWAYS be game/system audio
• Re-transcribe with correct naming

Problem: GPU not detected but I have NVIDIA GPU

Solution:

1. Install CUDA Toolkit from developer.nvidia.com/cuda-downloads
2. Update GPU drivers from nvidia.com/drivers
3. Restart your computer after installation
4. Verify in Task Manager > Performance > GPU

Problem: GPU doesn't have enough VRAM for Whisper model

Solution:

1. Use a smaller model (small instead of large)
2. Change compute_type from float16 to int8
3. Close other GPU-intensive applications
4. Switch to CPU processing as fallback

Need More Help?

Check the logs/ folder for detailed error messages.

Contact support: craig@nordiqai.io