Continue is the configuration-first option among open-source coding assistants. It plugs into VS Code and JetBrains IDEs and gives you three things — inline autocomplete, a chat panel, and edit commands — but the part that sets it apart is that every model slot is yours to fill.
You can point autocomplete at a small local model and chat at a frontier one, which is a practical way to keep sensitive code off a cloud provider while still getting good answers for hard questions.
Image: auto-generated social card of the continuedev/continue GitHub repository. (github.com/continuedev/continue)
What it covers
- Autocomplete with a model you choose, including local ones
- Chat with codebase context and referenced files
- Inline edit commands for refactors and documentation
- Shareable configuration so a whole team can use the same setup
What to know before you switch
Expect to spend time in the config file. That is the price of flexibility, and it is also why Continue is less popular with people who just want a plugin that works out of the box. Response quality is entirely a function of the model behind it.
Who it fits
Teams with a privacy policy to respect, developers with a local GPU, and anyone who wants the assistant’s behaviour to be reproducible and auditable rather than hidden behind a vendor default.