openmoto
commented
6 days ago Same issue here, never been able to get ollama external server working in any installation method.
OS: [Ubuntu Linux with docker, Ollama on external linux server] Browser: [Chrome, Brave, Edge]
I'm able to log into the container using docker exec -it bolt-bolt-ai-1 bash
Check the environment variable for ollama
printenv | grep OLLAMA_API_BASE_URL
the output shows it's set correctly
OLLAMA_API_BASE_URL=http://172.25.10.11:11434
yet, bringing the container up shows it's not set:
visvapravin
dustinwloring1988
Describe the bug
The current repo isn't working with Docker, Ubuntu Linux 22.04, with an External Ollama server.
The following is how to reproduce it: git clone https://github.com/coleam00/bolt.new-any-llm.git cd bolt.new-any-llm cp .env.example .env.local nano .env.local (edit the following line) OLLAMA_API_BASE_URL=http://192.168.50.84:11434 (when i do a curl from the docker host machine to 192.168.50.84:11434, I get 'ollama running' docker build . --target bolt-ai-production docker-compose --profile production up
The site will come up, but when I select Ollama in the drop down it doesn't have any of the models that I have installed populated.
I have also tried everything from switching the docker-compose.yaml line from RUNNING_IN_DOCKER=true to false.
Link to the Bolt URL that caused the error
http://192.168.50.38:5173/
Steps to reproduce
git clone https://github.com/coleam00/bolt.new-any-llm.git cd bolt.new-any-llm cp .env.example .env.local nano .env.local (edit the following line) OLLAMA_API_BASE_URL=http://192.168.50.84:11434 (when i do a curl from the docker host machine to 192.168.50.84:11434, I get 'ollama running' docker build . --target bolt-ai-production docker-compose --profile production up
Expected behavior
Should have a list of Ollama models that are displayed and that can be used
Screen Recording / Screenshot
No response
Platform
Additional context
No response