KRASH listens to your Klipper printer over Moonraker and persists byte offset, layer, physical Z and Z-offset every 3 seconds in the cloud, not on your SD card. When the MCU drops, your last known good state is one command away.
KRASH is intentionally small. It does one thing: keep enough state on a server you can reach so a hard MCU failure doesn't cost you 14 hours of print time.
One include line. Restart Klipper. The plugin registers a 3-second timer that snapshots virtual_sdcard + toolhead state.
Snapshot is POSTed to your account. Redis absorbs the high-frequency writes; Postgres only sees terminal states. SD card untouched.
Heartbeat stops at T+0. Server flags the print as recoverable after 90s of silence. Last good state stays warm in cache for 7 days.
Open the printer card. Copy the generated SET_KINEMATIC_POSITION, home X/Y, heat, then RESUME_AT the saved byte offset.
| Specifications all prices USD / mo | Local unlimited · client-side $0/forever | Starter 2 machines $5/mo | Pro 10 machines $15/mo | Enterprise 50 + $1/extra $49/mo |
|---|---|---|---|---|
| Cloud RAM persistencewrites never touch SD | - | ● | ● | ● |
| Disaster recovery syncresume after MCU loss | local-only | ● | ● | ● |
| SD card wear3-second writes | high | none | none | none |
| Heartbeat interval | 3.0s | 3.0s | 1.5s | 1.0s |
| Recovery cache window | - | 72h | 7d | 30d |
| Webhook + email alerts | - | both | both + SMS | |
| Print history retention | local | 30d | 1y | unlimited |
| Support | community | email · 48h | priority · 12h | dedicated |