What it is and who it's for
Continue is an open-source AI assistant for VS Code and JetBrains: inline autocomplete, chat about your code, edits and custom assistants — all running on whatever model you like, cloud or local. For developers who want a flexible Copilot-style helper without being locked to one provider, with a private local option. It stands out for covering both IDE ecosystems and for how deeply you can customize it through config.
Key features
- Inline autocomplete (tab) and chat about selected code.
- Any model: API or local (Ollama, LM Studio).
- Custom assistants, slash commands and context providers.
- Works in VS Code and every JetBrains IDE.
- @-context: files, docs, the whole repo.
- Open-source and free (you only pay for tokens if you go cloud).
Get started in 5 minutes
- Install the Continue extension in VS Code, or the plugin in JetBrains.
- Pick a model: plug in an API key or point it at a local one via Ollama.
- Accept completions as you type and open chat (Cmd/Ctrl+L) to ask about your code.
When to pick it, when not to
- ✅ Pick it if you want a flexible Copilot-style assistant where you choose the model, especially in JetBrains.
- ✅ Pick it if a private, local autocomplete and chat matter to you.
- ❌ Skip it → go with "Cline"/"Kilo Code" if you need an autonomous agent rather than a completion assistant.
- ❌ Skip it → go with "Cursor" if you want the most polished inline UX out of the box with zero setup.
The honest price
Free and open-source; you pay for tokens if you use a cloud model (local costs nothing). There are team and hub features too. Exact numbers are on the vendor's site and do change.
Gotchas
- Completion quality depends on the model — a local one is weaker than cloud.
- The flexibility comes with config work to tune it to your taste.
- It's an assistant, not a full autonomous task agent.
- Local autocomplete taxes your hardware (you need a decent GPU for it to feel fast).
🤖 Prompt accelerator
"I'm setting up Continue in <VS Code / JetBrains> on a
project. Help me configure it: which model to use for autocomplete and which for chat (cloud vs local Ollama on my GPU), and which context providers and slash commands to add for my workflow. Give me a sample config and the scenarios where the assistant will save me the most time."