Working with AI

Local models

Ollama

Can interact with VSCodium, both open source

Setting Up Ollama

Install Ollama (Free Local AI) and Download a free model

curl -fsSL https://ollama.com/install.sh

Install AI model llama3.2 (or other) in smallest size

ollama pull llama3.2:1b 

Verify the model appears in API

curl http://localhost:11434/api/tags

Test a simple chat

curl http://localhost:11434/api/chat -d '{

  "model": "llama3.2:1b",

  "messages": [{"role": "user", "content": "Hello, are you working?"}],

  "stream": false

}'

or

curl http://localhost:11434/v1/chat/completions \

  -H "Content-Type: application/json" \

  -d '{

    "model": "llama3.2:1b",

    "messages": [{"role": "user", "content": "Hello"}]

  }'

Quick Interactive Test

ollama run llama3.2:1b

Connect to VSCodium

You need to install extension called Continue. Then CTRL/L launches the AI sidebar. From there, you an open Models and VSCodium should detect your running agents.

Configure Continue to Use Your Ollama Model

There are a few ways; the simplest is manual config for full control.

1. In VSCodium, click the Continue sidebar icon (🧠).

2. At the top of the Continue panel, click the gear icon ⚙️ → Open config.json (or sometimes it says "Configure" / "Open Settings").

   • This opens ~/.continue/config.json (or config.yaml in newer versions — either works; many guides now use YAML).

3. Replace or add to the file with something like this (minimal setup for your model):
   Using JSON format (most common):
   JSON

{

  "models": [

    {

      "title": "Llama 3.2 1B (local)",

      "provider": "ollama",

      "model": "llama3.2:1b",

      "apiBase": "http://localhost:11434"

    }

  ],

  "tabAutocompleteModel": {

    "title": "Llama 3.2 1B Autocomplete",

    "provider": "ollama",

    "model": "llama3.2:1b",

    "apiBase": "http://localhost:11434"

  }

}

4. Or using YAML format (newer/recommended in some docs):
   YAML

name: Local Setup

version: 0.0.1

schema: v1

models:

  - name: Llama 3.2 1B Chat

    provider: ollama

    model: llama3.2:1b

    apiBase: http://localhost:11434

tabAutocompleteModel:

  name: Llama 3.2 1B Complete

  provider: ollama

  model: llama3.2:1b

  apiBase: http://localhost:11434

5. Save the file.

6. Continue may prompt you to reload / restart the extension — do so (or press Ctrl+Shift+P → "Developer: Reload Window").

Step 4: Use It!

• Chat: In the Continue sidebar, select your model from the dropdown at the top. Highlight code → right-click → "Ask Continue" (or press Ctrl+L / Cmd+L). Or just type in the chat box.

• Autocomplete: Start typing code → suggestions appear inline (accept with Tab). You can toggle it in settings if needed.

• Switch models: If you pull more (e.g. ollama pull qwen2.5-coder:3b), add them to the models array the same way — they'll appear in the dropdown.

• Test it: Ask something simple like "Explain this function" on selected code, or "Write a Python function to sort a list".

Tips & Troubleshooting

• Model not appearing / connection error?

  • Check Ollama is running: curl http://localhost:11434/api/tags should list models.

  • Restart VSCodium.

  • In Continue config, add "apiBase": "http://127.0.0.1:11434" if localhost has DNS issues (rare).