How to connect a local LLM with Ollama
You can run an AI model entirely on your own machine and use it inside Eclipse by serving it with Ollama and pointing the IDE at the local endpoint. Nothing leaves your device, no API key is required, and it works offline. Here's the short version.
What you need
A computer that can run a local model, Ollama installed, and Eclipse. Local models trade some capability for privacy and zero usage cost, so pick a model size your hardware can handle comfortably.
Step 1 — Install Ollama and pull a model
Install Ollama, then pull a model of your choice. Ollama serves an OpenAI-compatible API locally, usually at http://localhost:11434. Once the model is downloaded, it runs on your machine.
Step 2 — Point Eclipse at the local endpoint
In Eclipse's AI settings, add a provider of type Ollama (or a generic OpenAI-compatible endpoint), set the URL to your local Ollama address, and select the model you pulled. No key is needed for local models.
Step 3 — Use it (and optionally add tools)
Open the AI chat and start working with your local model. To let it act on your files or run commands, enable MCP so agents can call real tools — still fully local if your tools are local.
Why run locally?
Local inference keeps prompts and source on your device, needs no internet, and has no per-token bill. That makes it a strong fit for regulated, offline or high-volume work. More on this in local & private AI.
FAQ
Do I need an API key for a local model?
No. Local models served through Ollama need no key. You only add a key when you connect a cloud provider such as OpenAI or Claude.
Will my code be sent anywhere?
No. With a local model, prompts and code are processed on your own machine and are not sent to a third-party provider.