Skip to content

SETUP.md

Latest commit

 

History

History
166 lines (122 loc) · 3.22 KB

SETUP.md

File metadata and controls

166 lines (122 loc) · 3.22 KB

Quick Setup Guide

1. Install Ollama and Models

Download and install Ollama from https://ollama.ai

Then install the models:

ollama pull gemma3:4b
ollama pull gemma3:8b
ollama pull deepseek-r1:8b

Verify installation:

ollama list

2. Start the Python Server

Open a terminal in the project folder:

cd server
pip install -r requirements.txt
python server.py

You should see:

Starting Local AI Copilot Server...
Available models: ['gemma3-4b', 'gemma3-8b', 'deepseek-r1-8b']
Server running on http://127.0.0.1:5000

Keep this terminal running!

3. Build the VS Code Extension

Open another terminal:

cd vscode-extension
npm install
npm run compile

4. Run the Extension

In VS Code:

  1. Press F5 (or Run → Start Debugging)
  2. A new VS Code window will open with the extension loaded
  3. Press Ctrl+Shift+P and type "Local AI Copilot: Chat"
  4. Start chatting!

Testing the Setup

Test 1: Server Health Check

Open browser to: http://127.0.0.1:5000/health

Should show:

{"status":"healthy","ollama":"running"}

Test 2: Ollama Direct

ollama run gemma3:4b "Write a hello world in Python"

Test 3: Server API

$body = @{
    prompt = "Write hello world in Python"
    model = "gemma3-4b"
    task_type = "code"
} | ConvertTo-Json

Invoke-RestMethod -Uri "http://127.0.0.1:5000/completion" -Method POST -Body $body -ContentType "application/json"

Common Issues

Issue: "Failed to connect to AI server"

Solution: Make sure the Python server is running

cd server
python server.py

Issue: "Ollama API error"

Solution: Make sure Ollama is running

ollama list

If Ollama isn't running, start it:

  • Windows: Ollama should auto-start. Check system tray.
  • Mac: ollama serve
  • Linux: ollama serve

Issue: "Model not available"

Solution: Pull the model

ollama pull gemma3:4b

Issue: Extension commands not showing

Solution:

  1. Close the debug VS Code window
  2. In your main VS Code window, run: 
    cd vscode-extension
npm run compile
  3. Press F5 again

Using the Extension

Chat Command

  1. Ctrl+Shift+P → "Local AI Copilot: Chat"
  2. Select model from dropdown
  3. Type your question and press Enter

Code Editing

  1. Select some code
  2. Right-click → "AI: Edit Code"
  3. Describe what you want to change
  4. Review the diff and apply

Code Generation

  1. Place cursor where you want code
  2. Right-click → "AI: Generate Code"
  3. Describe what you want to create
  4. Code will be inserted

Code Explanation

  1. Select code you want explained
  2. Right-click → "AI: Explain Code"
  3. Read explanation in chat panel

Tips

  • 💡 Use gemma3-4b for quick responses
  • 💡 Use deepseek-r1-8b for complex reasoning tasks
  • 💡 Select code before asking questions for better context
  • 💡 The chat panel remembers your conversation
  • 💡 Code suggestions show "Before/After" diffs

Next Steps

Once everything works:

  1. Package the extension: vsce package
  2. Install it: code --install-extension *.vsix
  3. Stop using F5 and use it like a real extension!

Need Help?

Check the full README.md for detailed documentation.