Tutorials · Chapter C (3/4) · ~8 min

Local LLMs & Ollama

Try it → see it → read → next

Run models on **your machine** for privacy and offline use — with different tradeoffs than cloud APIs.

On paths: Builder

Optional on: Student

Try tools: Open source & local

Try yourself

Playground

Local vs cloud LLM

Match each scenario to local (Ollama-style) or cloud API — tradeoffs, not religion.

Summarize patient notes on a hospital laptop — data cannot leave the device.

Recap

What you just did

You matched scenarios to local vs cloud — privacy/offline vs scale/frontier models.

Teach

How it works

| | Cloud API | Local (e.g. Ollama) | |---|-----------|---------------------| | Privacy | Data sent to provider | Stays on device | | Model size | Frontier models | Smaller open weights | | Cost | Per token | Hardware + electricity | | Setup | API key | Install + download weights |

Neither is always right — many products use both (local draft, cloud for hard tasks).

Read

Try it

Open Ollama or your team's approved local runner. Pull one small model and run a offline prompt. Compare latency and quality to your usual cloud chat.