Expanding Detection Mode

Not every code change should be reviewed the same way.

Some reviews should prioritize broad bug detection, even if that produces some lower-stakes findings. Other reviews should surface only the highest-confidence and severe issues. Routine changes in less sensitive parts of the codebase may call for a fast, economical review. For the most consequential changes, avoiding a single critical bug may justify spending dramatically more on a demonstrably more exhaustive review.

These priorities involve real tradeoffs among recall (bug detection), precision (false positives), speed, and cost. No single configuration maximizes every dimension. AI code review should make these tradeoffs explicit and let you choose the right balance for each change.

That's why we're expanding Detection Mode, allowing you to choose what performance profile you want for your code reviews:

• Budget — Fast, low cost bug detection. Lower bug detection and precision compared to more premium modes.
• Balanced (default) — Industry leading high bug detection. Recommended for teams who use Agents to review PR comments, where the inconvenience of occasional false positives is minimal.
• Precise — Similar bug detection as Balanced mode, but higher confidence findings due to additional agentic validation to reduce noise from false positives and low-value findings. Recommended for teams who prefer to spend more budget (dollars and latency) to ensure high precision findings.
• Ultra — Frontier bug detection for the most consequential reviews. Leverages a much higher number of parallel agents using more advanced coding models. Useful when maximizing bug detection matters more than cost.

There are a few ways to configure which Detection Mode is used during a review:

• At the repository level (the default is Balanced). Configure in Settings → Repos.
• At the personal GitHub author level (by default, the Detection Mode configured on the repository of the PR is used). Configure in Settings → Personal.
• On an ad hoc review basis by mentioning @macroscope-app in an issue comment (e.g. “@macroscope-app do an Ultra review”).
• On an ad hoc PR basis, by applying one of a few magic PR labels (e.g. macroscope-ultra, macroscope-balanced, macroscope-precise, macroscope-budget). All subsequent pushes to that PR will use the labeled mode.

These modes are not simply "better" and "worse." Each is purpose-built for a different performance profile—and a price that reflects the compute and complexity required to run it.