Continue.dev

Continue is an open-source AI code assistant that gives developers complete control over which language models power their AI coding experience. It supports VS Code and JetBrains IDEs and can connect to any LLM provider, from cloud services like OpenAI and Anthropic to locally running models via Ollama and LM Studio, making it the most flexible AI coding tool available. Continue's core value proposition is provider independence. While tools like GitHub Copilot and Cursor are tied to specific AI providers, Continue lets developers choose and switch between any combination of models for different tasks. You might use Claude for chat-based reasoning, GPT-4 for code generation, and a fast local model for tab autocomplete, configuring each independently. This flexibility extends to self-hosting: organizations can run Continue with models deployed on their own infrastructure, ensuring that no code ever leaves their network. The tool provides tab autocomplete, an in-editor chat interface with codebase context, and inline code editing through slash commands. Custom context providers allow developers to connect additional knowledge sources like documentation, issue trackers, or internal wikis to their AI assistant. The configuration is handled through a straightforward JSON file, making it easy to share setups across a team or maintain different configurations for different projects. Continue is ideal for developers who want control over their AI tooling stack, whether for privacy reasons, cost optimization, or simply preferring a particular model's coding style. It is especially valuable for teams in regulated industries that need to self-host their AI infrastructure, and for developers who want to experiment with different models to find the best fit for their programming language and workflow. The open-source nature means there is no vendor lock-in and the community actively contributes context providers, slash commands, and configuration templates. Continue is completely free and open-source. The only cost is whatever LLM API usage you incur, which can be zero when using local models. The tradeoff for this flexibility is a less polished out-of-the-box experience compared to commercial alternatives. Initial setup requires configuring your preferred models and providers, and the quality of the experience depends heavily on which models you choose. The user interface and feature set, while actively improving, are not yet as refined as Cursor or Copilot.