From Extension to Mission Control

Continue started as an open-source IDE extension, but as teams began using coding agents more deeply, the focus shifted. Now the question wasn't how fast you could write code, but how much you could trust the changes inside a PR.

As PRs got bigger and faster, the number of items that needed repeated checking also grew. Instead of manually re-checking things like basic security checks, documentation sync, missing metrics, and AI slop creeping in every time, the team documented these standards as something called AI Checks.

AI Checks are small, single-purpose agents, each defined by a single markdown file. They run as GitHub status checks on every PR, returning green if the code meets the standard, or red along with a suggested diff if it doesn't. The key is not adding new opinions, but simply judging whether the code violates standards that have already been set.

Based on this structure, the center of gravity for Continue has also shifted.

• Continue CLI (cn): An open-source coding agent that runs AI Checks, usable in a terminal, CI, or any environment you want.
• Mission Control: An orchestration layer that coordinates multiple cloud agents in parallel, attaching team permissions and integrations to run checks on every PR.
• AI Checks: The entry point where teams define what counts as "good code" in their codebase using markdown, automatically verified on every PR.

The existing VS Code extension will focus on stabilization, and the JetBrains plugin will transition to community-maintained. Instead, the CLI becomes the current center of active development, Discord is being shut down, and the hub for support and discussion is moving to GitHub Discussions and r/ContinuousAI.

The message is clear. AI isn't automation that replaces people—it produces the best results when it's more tightly attached to the human feedback loop. Continue started as the first feedback loop inside the editor, and is now moving to the next stage: automatically enforcing a team's standards.