Most AI coding tools make a bet for you: one vendor's model, one closed extension, one way of working. Continue's bet is the opposite — it treats the IDE assistant as configuration, not a product, so the same setup drives autocomplete, chat, inline edit, and agent mode against whatever model you choose. That choice is why it spread to 34k+ GitHub stars and why Cursor eventually acquired the team in June 2026.
What Sets It Apart
- Model-agnostic by design: a single YAML-style config maps each capability (tab-complete vs. chat vs. agent) to a different provider, so you can run a fast local model for autocomplete and a frontier model for reasoning in the same session.
- Four modes, one surface: autocomplete, chat, edit, and agent share context and history instead of living in separate tools, which keeps codebase context consistent as you move between asking and acting.
- Built to be customized and shared: assistants are defined as portable artifacts a team can version and reuse, rather than per-developer settings buried in an editor.
Who It's For
Great fit if you want to avoid vendor lock-in, run local or self-hosted models, or standardize a team's AI setup as shareable config. Look elsewhere if you'd rather have a polished, opinionated turnkey experience — and note the original repo is now archived and read-only following the Cursor acquisition, so weigh the project's future direction before committing to it long-term.
