tool·AI Tools

Continue

Open-source AI code assistant extension for VS Code and JetBrains—BYOK or local.

FREEOPEN SOURCEBYOKLOCAL-FIRST

Setup guide · 25–35 min

Visit Continue →

Quick facts

Price model: Open source
Starting price: Free
Best for: IDE-integrated AI · Local model coding · Custom model routing
Replaces: GitHub Copilot, Cursor, Codeium Pro
Platforms: MacWindowsLinux
Last verified: 2026-06-22

Why it's listed

Keep your editor and swap AI backends freely—no lock-in to one copilot vendor.

Continue adds chat, autocomplete, and edits inside your editor using Ollama, OpenAI, or other providers. Avoid a second IDE subscription when you can wire your own models and keys.

The catch

Jump to setup guide ↓

Setup and model tuning take time; autocomplete quality varies by model.

How to set up Continue

AI chat and edits inside VS Code (or JetBrains) using Ollama locally or your own API keys—no Copilot subscription required.

Time: 25–35 min
Difficulty: Moderate
Verified: 2026-06-22

Before you start

VS Code or JetBrains IDE installed
Ollama running with at least one code-friendly model (e.g. qwen2.5-coder:7b), or an OpenAI/Anthropic API key
A project folder open in the editor

Install the extension
VS Code: Extensions → search Continue → Install. JetBrains: Plugins → Continue. Reload the editor when prompted.
Open Continue and sign in locally
Click the Continue icon in the sidebar. First run walks you through onboarding—you can skip cloud signup and stay local-only.
Add Ollama as a model
Continue config (JSON in sidebar gear) → models. Add Ollama provider pointing at http://localhost:11434 with a pulled coder model. Set it as default chat model.
Optional — add a cloud model (BYOK)
In the same config, add OpenAI or Anthropic with your API key. Use for harder refactors; keep Ollama for unlimited free tries.
Chat with your codebase
Highlight a function → Ask Continue to explain or refactor. Use @ to reference files in context. Accept diffs inline like a normal review.
Tune autocomplete (optional)
Enable tab autocomplete in Continue settings. Smaller local models autocomplete faster; larger models read more context but need more RAM.

Troubleshooting

Ollama connection refused: Run ollama serve (or restart Ollama app). Confirm curl http://localhost:11434/api/tags works.
Suggestions are nonsense: Model too small for your repo. Pull a coder-tuned tag or switch chat to a cloud model for that task.
Extension won't load config: Invalid JSON in config.json—Continue shows line errors. Start from the default template in docs.

Keep it working

Pull updated Ollama coder models when releases land
Rotate API keys in config if you use BYOK—never commit keys to git
Update the extension when VS Code prompts; breaking changes are rare but noted in changelog

Official docs: docs.continue.dev/setup/overview