ollama CLI
ollama serve
- You should run it to start the ollama server, it's necessary for all API operations (e.g., pull, run, etc)
- You can also autostart it on login:
brew services start ollama
ollama list
- List all the downloaded models
ollama ps
- List running and exposed models
ollama pull
- Download a model only
- Models are saved at
~/.ollama/models/
- After a model is pulled, it's ready to be used via API
ollama pull nomic-embed-text
ollama run
- Download, run a model and plug it to stdin
- If you want just to download ->
ollama pull
ollama run gpt-oss
ollama run gemma3:270m
# same! (the model must be pulled first)
curl http://localhost:11434/api/chat \
-d '{
"model": "gpt-oss",
"messages": [{"role": "user", "content": "Hello!"}]
}'
ollama stop
ollama rm