From Extension to Mission Control

We started Continue as an open-source IDE extension because that's where the work was, right next to your code, in your editor. Hundreds of thousands of you adopted it. That still means everything to us.

But over the past year, something shifted. As our own team leaned harder into coding agents, we stopped worrying about writing code and started worrying about trusting it. PRs got bigger and faster than our review bandwidth. We kept re-checking the same things: security basics, docs staying in sync, metrics not silently breaking, AI slop sneaking through. The anxiety wasn't "can I read every line" but "did I miss the one thing that matters."

So we started writing those concerns down as AI Checks: small, single-purpose agents, each one just a markdown file, that run on every PR as GitHub status checks. Green if the code looks good, red with a suggested diff if not. No comment walls. No unsolicited opinions. Just: does this PR violate something you've already decided matters?

That idea became the core of what Continue is now.

What's new

Continue CLI (cn) is an open-source coding agent that powers AI Checks on Mission Control. It can also be used in your terminal, CI, or any environment of your choice.

Mission Control is the orchestration layer on top: cloud agents working in parallel, running checks on every PR, with team permissions and integrations. Same relationship Git has to GitHub. The open-source tool does the work, the platform coordinates it across a team.

AI Checks are where most teams will start. Define what "good" looks like for your codebase in markdown. Continue enforces it on every PR. To generate suggested checks for your repo, run this in Claude Code or your preferred agent:

Help me write checks for this codebase: <https://continue.dev/walkthrough>

It'll explore your codebase, interview you, and write checks to .continue/checks/, which you can run on every PR.

What's changing

VS Code extension is stable and polished. That's what you've been asking for, and it's a priority.

JetBrains plugin is now community-maintained. We know this is disappointing for some of you. IDE extensions are hard to maintain across completely different frameworks, and we had to make honest choices about where a small team's effort creates the most value. We'll still review and merge PRs. If you're on JetBrains, we'd strongly recommend trying the CLI. It works in any environment and it's where our active development is happening.

Discord is being retired. Community support has moved to GitHub Discussions. Broader conversation is at r/ContinuousAI. Everything stays discoverable and searchable.

The mission hasn't changed

Amplified, not automated, developers. That's been the line from the beginning, and it's not marketing. It's a design principle. AI tightly coupled to human feedback loops generates better outputs than AI running autonomously. We started with the first feedback loop: you and your editor. Now we're building the next one: your standards, enforced automatically, with you still deciding.

We're just getting started. We'd love your feedback.