Help with local ollama setup?

[jasdjensen]

I've tried to setup a connection to my local ollama. I can select a model using http://192.168.1.106:11434/v1 but can't get any further. I've tried putting in a dummy APIkey but that doesn't help.

This works from the docker host. (to test connectivity and my ollama instance is running)

curl -X POST http://192.168.1.106:11434/api/generate -d '{
  "model": "llama3.3:latest",
  "prompt": "How many letters are in the english alphabet?",
  "stream": false
}'

My log looks good:

Session validated for user: admin
Authentication successful, proceeding with request
Saving userData (0 TV recommendations, 0 movie recommendations)
Writing 1808 bytes to /app/server/data/user_data.json
Deleted existing /app/server/data/user_data.json
Successfully wrote new /app/server/data/user_data.json
✓ /app/server/data/user_data.json exists, size=1808 bytes, modified=Thu Mar 13 2025 04:11:33 GMT+0000 (Coordinated Universal Time)
✓ Read back file, content length: 1808 bytes
✓ Verified file contains properly encrypted data
✓ Decrypted data has 0 TV recommendations
✓ Decrypted data has 0 movie recommendations

Not saying there is a bug, but I'm having trouble. Webpage keeps returning to Settings / AI Service tab. I can select different models and save but can't get off that screen to run recommendations.

[TannerMidd]

Please see these comments for now:

#31 (comment)

#31 (comment)

[TannerMidd]

You can also try out the latest dev build that should fix the issue.

docker pull tannermiddleton/recommendarr:dev

You may have to clear your browser cache etc..but I can confirm it works now assuming you have Ollama setup to be available on the local network instead of only the loopback (127.0.0.1)

[jasdjensen]

Thank you. This helped.