Continue
The leading open-source AI code assistant
Continue is the leading open-source AI code assistant that works inside VS Code and JetBrains IDEs. Unlike proprietary tools, Continue lets you bring your own models (BYOK). You can connect to local LLMs via Ollama, or use API keys for GPT-4, Claude 3.5, and DeepSeek. It gives you full control over your data privacy and the AI stack you use.
Category
Extension
Rating
Users
400K+
Platform Support
Pricing Plans
Free Plan
$0
Free & Open Source (Apache 2.0)
Pro Plan
Individual: Full features, bring your own API key or run local models
Pro Plan
Enterprise: Hub for centralized configuration, analytics, and governance
Enterprise Plan
定制
Continue for Teams: SSO, audit logs, standardized config
Core Features
1 Core Features
Model Agnostic
Use any model you want: GPT-4o, Claude 3.5 Sonnet, DeepSeek, or local Llama 3
Context Providers
Type @ to reference files, docs, terminal output, or git diffs in your chat
Local AI Support
First-class support for Ollama, LM Studio, and other local inference servers
Tab Autocomplete
Predictive code completion using low-latency models like StarCoder2 or DeepSeek-Coder
2 Flexibility
Customizability
Configure everything via config.json: custom prompts, slash commands, and model parameters
Cross-IDE
Consistent experience across VS Code and JetBrains (IntelliJ, PyCharm, etc.)
3 Developer Experience
Edit Mode
Highlight code and press Cmd+I to refactor or fix bugs inline
Docs Indexing
Index documentation sites to ask questions about third-party libraries
Data Privacy
Your code never leaves your machine unless you use a cloud API; full telemetry control
Pros and Cons
Advantages
- Completely open-source and free to use
- Unmatched flexibility: swap models instantly (e.g., use DeepSeek for chat, StarCoder for autocomplete)
- Best-in-class support for Local LLMs (Ollama)
- Powerful context system (@docs, @codebase, @git)
- No vendor lock-in; you own your configuration
Disadvantages
- Requires more initial setup (configuration) than Copilot
- Autocomplete latency depends on your chosen provider/hardware
- UI is functional but less "native" feeling than Cursor
- Indexing for @codebase can be slower on very large repos compared to specialized vector DBs
Best For
Privacy-conscious developers who want to run local models
Power users who want to mix and match different AI models
Teams that need a self-hosted AI coding solution
Developers using JetBrains IDEs who want a better chat experience
Tech Stack
Architecture
IDE Extension (VS Code & JetBrains)