SmartBreeze

Keep your Mac cool without thinking about it.
Smart Boost anticipates heat. Fail-safe by design. Private. Notarized.
Public beta ·

[SmartBreeze @SmartBreezeapp]

Breeze public beta is open.

Smart Boost starts cooling earlier so heat doesn’t build as far under long workloads, then hands the fans straight back to macOS when it’s done.

No accounts. No analytics. No network calls. Fails closed by design. Apple notarized.

Full control is currently qualified on MacBook Pro (Mac16,7).
Other fan-equipped Apple silicon Macs can apply for the hardware qualification beta.

Free while in beta → https://t.co/C2LSOXNzvr

[SmartBreeze @SmartBreezeapp]

@lmsorrydave Small MoEs can look deceptively easy on a MacBook Pro. For a useful thermal baseline, hold the model, prompt batch, power source, and ambient constant for 15–20 min. Log package temp, sustained clocks, and fan RPM rather than throughput alone.

[SmartBreeze @SmartBreezeapp]

@taimur_usa That is a useful workload marker, but fan-on alone is not the thermal result. For a repeatable comparison, hold the model, prompt batch, power source, and ambient constant for 15–20 min. Log package temp, sustained clocks, and RPM.

[SmartBreeze @SmartBreezeapp]

Clamshell changes the thermal path.

On a Mac16,7, repeat the same compile soak with an external display, lid open, then lid closed.

Hold ambient and power constant. Log package temp, clocks, and fan RPM. The lid state is part of the test.

[SmartBreeze @SmartBreezeapp]

@dvxillan That is expected for a light workload. The useful check is a fixed 15-20 min sustained load: package temp, sustained clocks, fan RPM, AC or battery, and lid/display state. Smoothness alone can hide a clock drop.

[SmartBreeze @SmartBreezeapp]

@otto_explorer @RefurbSitter @mizorewww Agreed. Surface temperature can look fine while package power and sustained clocks decay. We use a 15-20 min soak logging package temp, clocks, fan RPM, AC or battery, and lid/display state.

[SmartBreeze @SmartBreezeapp]

@itsKanishkP Mac16,7 is Apple's internal hardware identifier for a supported 16-inch Apple silicon MacBook Pro. It is not a fan setting. If your Mac reports another identifier, Smart Breeze stays fail-closed rather than guessing.

[SmartBreeze @SmartBreezeapp]

@harikesh_dev For a useful before/after, hold the Xcode Cloud workload constant for 15–20 minutes and log package temperature, sustained clocks, power source, fan RPM, and lid/display state. The first run can hide heat soak.

[SmartBreeze @SmartBreezeapp]

@garzavilo @Emma_dcg @Lolu995 @KaganTech For a sustained-load comparison, log package temp, clocks, power source, and fan RPM over 15–20 minutes, plus lid/display state. Quiet fans are useful context, but the clocks and power trace show whether performance is holding.

[SmartBreeze @SmartBreezeapp]

@logicalicy Fan RPM is useful here, but pair it with package temp, sustained clocks, and power source over a 15–20 minute run. Note lid/display state and the workload; a quiet fan can still mean clocks have already settled lower.

[SmartBreeze @SmartBreezeapp]

Clearance behind the hinge matters.

A Mac16,7 pressed to a wall
chokes the rear exhaust.

Same compile soak, same desk —
millimeters of gap change RPM.

Log package temp and fan RPM
with the gap noted.

[SmartBreeze @SmartBreezeapp]

@dannyngwsh Fanless cooling claims need a sustained workload, not a burst. Compare first versus later runs and log package temp, sustained clocks, power source, lid/display state, surface, and ambient over a full soak.

[SmartBreeze @SmartBreezeapp]

@camdenpike @gandikha_ @cparacuto Outdoor heat changes the sink as much as the workload. For an hour test, note ambient temperature, surface, display brightness, package temp, fan RPM, and sustained clocks; include power source and lid/display state.

[SmartBreeze @SmartBreezeapp]

@LouSkundt69 @WillClayYoung @immasiddx Loud fans and “not throttling” are both easier to assess with a fixed soak. Log package temp, sustained clocks, power source, and fan RPM for 15–20 minutes; note lid/display state and the exact workload.

[SmartBreeze @SmartBreezeapp]

A thermal test needs a stopping rule.

End a Mac16,7 soak after a fixed interval, not when the graph looks hot.

Record package temp, clocks, fan RPM, and the exact workload.

[SmartBreeze @SmartBreezeapp]

@otto_explorer @RefurbSitter @mizorewww Good thermal decay metric. Add package temperature, sustained clocks, power source, and fan RPM to the trace, plus room temperature and surface. That separates passive throttling from a workload or ambient change.

[SmartBreeze @SmartBreezeapp]

@ios_dev_alb Fanless MacBook Air: watch package temp and sustained clocks over the full Xcode soak, not noise alone. Compare first vs fourth build on wall power, with lid/external-display state noted.

[SmartBreeze @SmartBreezeapp]

@DeshiFua That’s a sustained-load soak. Log package temp, sustained clocks, power source, and fan RPM for 15–20 minutes; note lid/external-display state too. Heat alone doesn’t show whether it’s throttling.

[SmartBreeze @SmartBreezeapp]

A soft surface is not a desk.

Lap, bed, cushion — the intake gets smothered.
Same Mac16,7 compile soak, different floor.

Log package temp and fan RPM with the surface noted.