USE GUIDE — OLLAMA

Main takeaway: start local chat first, then test API, then automate

API means application programming interface, a URL apps use to exchange data with your local model.

WINDOWS

Model check

ollama list
ollama run llama3.2:3b "Say hello in one short sentence."

Expected result: model list appears and model responds in one sentence.

macOS

Model check

ollama list
ollama run llama3.2:3b "Say hello in one short sentence."

Expected result: same output pattern as Windows.

Linux

Model check

ollama list
ollama run llama3.2:3b "Say hello in one short sentence."

Expected result: local model answers directly in terminal.

API TEST

Call localhost API

curl http://localhost:11434/api/generate -d '{"model":"llama3.2:3b","prompt":"Give one productivity tip.","stream":false}'

Expected result: JSON response that includes generated text.

PRACTICAL USE CASE

Private knowledge draft assistant

Feed sanitized internal notes to a local model, generate draft replies, and keep sensitive details on your own machine. This is ideal for support teams, operations notes, and repetitive report writing.

TROUBLESHOOTING

Common fixes

If ollama list fails, start server with ollama serve.

If model is missing, run ollama pull llama3.2:3b.

If port 11434 fails, restart Ollama service and verify no firewall block.