Continue

Continue is an AI assistant and agent framework built around your development environment. It runs as an IDE extension (with documented support for VS Code, Android Studio, and code-server) and connects those editors to large language models and tools that can act on your codebase and environment. On top of the editor plugins, it offers a Mission Control web interface for managing cloud-based agents.

The product is aimed squarely at developers and engineering teams who want deep IDE integration plus centralized control over AI agents. From the docs, you configure models and behavior via local YAML/JSON config files and/or through Mission Control, including details like certificates, proxies, and provider-specific options. It supports popular model providers such as OpenAI and OpenRouter, as well as local and remote models through Ollama, and can run in remote/dev container setups like code-server and WSL.

According to the documentation, Continue supports an "agent mode" with tools, and a Mission Control "Agents" page that gives you a single view of all your cloud agents—where they run, how they’re triggered, and how to reuse them. This points to a genuinely agentic setup, not just inline chat: agents can be wired to external tools, depend on secrets and environment variables, and operate across projects. At the same time, this flexibility comes with configuration overhead: you may need to manage certificates, proxy behavior, local vs. remote Ollama connectivity, and per-model tool capabilities before everything works smoothly.

There are some clear technical caveats in the docs. Codebase indexing on Linux requires modern CPU features (AVX2 and FMA), and will be disabled otherwise. Models must explicitly support tools for agent mode to function, and a fair amount of troubleshooting guidance is devoted to networking issues (corporate proxies, SSL certificates, WSL, Docker) and provider-specific errors (OpenAI, OpenRouter, Ollama). Privacy and data-handling practices aren’t detailed in the FAQ beyond how secrets and environment variables are stored (locally via .env files or managed through Mission Control), so you’ll need to consult their main site or security docs if that’s critical.

Continue is a good fit if you’re a developer or team comfortable editing config files and debugging IDE/network quirks to get a powerful, customizable agent setup. It’s not well suited to non-technical users looking for a simple web chatbot, or to environments that can’t satisfy the documented hardware and networking requirements. Pricing information isn’t provided in the FAQ, so you’ll need to check their main site or repository for current licensing and plans.