Which AI models can I use with Continue?

You can use Anthropic, OpenAI, Ollama for local models, or any OpenAI-compatible endpoint. This flexibility allows you to standardize on a single workflow without committing to one vendor.

Is Continue difficult to set up?

Setup involves config files, model routing rules, and a learning curve. It is less polished than competitors like Cursor or Copilot, which deliberately hide these complexities.

What is the cost of using Continue?

The tool is free and open-source. However, you pay for the model costs associated with the API or local resources you choose to use.

Continue

✓ Editorially verified

Open-source, self-hostable VS Code/JetBrains AI assistant.

Free· Free / open-source; you pay model costsCodingBYO (any OpenAI-compatible API + Ollama for local)7.9 / 10

Visit website →

In short

Continue is an open-source AI assistant for VS Code and JetBrains that supports bringing your own model. It is best for teams requiring self-hosted solutions or local model support via Ollama to keep code private.

Best for

Pick Continue if you want a free, self-hostable AI coding assistant with local-model support.

Skip if

Skip it if you value zero-config UX over flexibility.

Continue is an open-source AI coding extension that runs in VS Code or JetBrains. You BYO model — Anthropic, OpenAI, Ollama for local, or any OpenAI-compatible endpoint — which makes it the natural pick for teams that want to standardise on a single AI workflow without committing to a single vendor.

Local-model support via Ollama is a quiet superpower. For privacy-sensitive work or air-gapped development, you can route everything through a local Llama or DeepSeek model and never send code to a third party.

The trade-off is polish. Setup involves config files, model routing rules, and a learning curve that Cursor and Copilot deliberately hide. Continue rewards teams who want control and punishes teams who want "it just works."

Editor's take

Continue is the answer to "can we run AI coding without sending code to a vendor?" — and the answer is yes, with effort. For the right team it's a strategic choice; for everyone else, Cursor's UX wins.

— The AI Tool Bible editorial team

Pros

✅ Free and open source
✅ Works with local models via Ollama
✅ No data leaves your machine if local
✅ BYO model flexibility

Cons

⚠️ Less polished than Cursor/Copilot
⚠️ Setup is fiddly

Use cases

self-hostedopen sourceBYO model

Frequently asked

Which AI models can I use with Continue?: You can use Anthropic, OpenAI, Ollama for local models, or any OpenAI-compatible endpoint. This flexibility allows you to standardize on a single workflow without committing to one vendor.
Does Continue support local models for privacy?: Yes, it supports local models via Ollama. This allows you to route everything through a local Llama or DeepSeek model, ensuring code never leaves your machine for privacy-sensitive work.
Is Continue difficult to set up?: Setup involves config files, model routing rules, and a learning curve. It is less polished than competitors like Cursor or Copilot, which deliberately hide these complexities.
What is the cost of using Continue?: The tool is free and open-source. However, you pay for the model costs associated with the API or local resources you choose to use.