Continue + Ollama: Fully Local AI Coding in VS Code
Set up Continue and Ollama for private, fully local coding in VS Code with this 2026 guide.
Table of Contents
- Introduction
- Prerequisites & Verification
- Configuring Continue to Use Ollama
- Connecting & Testing
- Daily Usage in Your Repo
- Build a Hello World App
- Best Practices & Pro Tips
- Troubleshooting
- Next Steps & Advanced Usage
Already installed? You’re 10 minutes away from unlimited local AI coding.
FULLY LOCAL AI POWERHOUSE
Continue.dev + Ollama:
Your Complete Local AI Coding Setup in VS Code
Welcome! You’ve already installed Continue.dev and Ollama is running. This guide shows you exactly how to turn VS Code into a fully private, unlimited, high-performance AI coding assistant that rivals (and often beats) expensive cloud tools like Augmentcode AI — with zero monthly fees and zero data leaving your computer.
100% offline after setup
Unlimited context & usage
Works exactly like Augmentcode
Why this matters in 2026: Modern open-source models (Qwen2.5-Coder, DeepSeek-Coder-V2) now match or exceed Claude 3.5 Sonnet on coding tasks — all running locally on your GPU.
02 Prerequisites & Verification
-
1Ollama is running
Open terminal and run:ollama list
You should see your downloaded models.
PS _> ollama list NAME ID SIZE MODIFIED qwen2.5-coder:7b dae161e27b0e 4.7 GB 6 hours ago PS _>
-
2Continue.dev extension is installed
Look for the Continue icon in the VS Code Activity Bar (left sidebar).
03 Configuring Continue to Use Ollama
Continue stores its configuration in a single JSON file.
-
1. Open the config file
Press Ctrl + Shift + P, type
Continue: Open config.jsonand press Enter.
[Screenshot: VS Code command palette showingContinue: Open config.json]
-
2. Replace with this recommended 2026 config
{ "models": [ { "title": "Qwen2.5-Coder 32B", "model": "qwen2.5-coder:32b", "provider": "ollama", "apiBase": "http://localhost:11434" } ], "tabAutocompleteModel": { "title": "Qwen2.5-Coder 14B", "model": "qwen2.5-coder:14b", "provider": "ollama" }, "contextProviders": [ { "name": "codebase" }, { "name": "terminal" } ] }
04 Connecting and Testing the Setup
- Click the Continue icon in the sidebar
- In the chat, type
/modeland select your Ollama model - Ask: “Hello, are you running locally?”
You should get an instant local response
[Screenshot: Continue chat window showing Ollama model selected and local response]
05 Using Continue in Your Repository
Open any Git repo in VS Code. Continue works exactly like Augmentcode AI.
Common Workflows
- Highlight code → Right-click → Continue: Edit
- Highlight + Ctrl + L → “Refactor this”
- Sidebar chat: “Add error handling”
- Use
@codebasefor full repo context
06 Build a “Hello, World” Application – Full Tutorial
We’ll build a modern Python FastAPI backend.
Step 1: Create folder and open in VS Code
mkdir fastapi-hello && cd fastapi-hello
Step 2: In Continue chat, type:
"Create a complete FastAPI project with main.py, requirements.txt, and a /hello endpoint"
07 Best Practices & Pro Tips
- Be specific in prompts: “Refactor this to be async with proper error handling”
- Use
@codebasefor large projects - Use smaller models (14B) for autocomplete, larger (32B) for complex tasks
- Keep Ollama running in the background
08 Troubleshooting
Ollama not connecting? Run
ollama serve in a separate terminal.
Slow responses? First token is slower; subsequent ones are fast. Use a smaller model if needed.
09 Next Steps & Advanced Usage
- Try multiple models in parallel
- Enable Continue Agents
- Install MCP for deeper context
- Create custom slash commands
You now have a world-class local AI coding setup.
Enjoy unlimited coding — share your wins on X @RealKyleRode
Enjoy unlimited coding — share your wins on X @RealKyleRode