PelletPilot

Three ways to run it

Local, self-hosted, or public — your choice

🏠 Local

The app talks straight to your grill on your LAN. No account, no cloud, no vendor. Lowest latency.

🔒 Self-hosted

Run the PelletPilot server yourself. Reach your grill from anywhere, on infrastructure you own.

☁️ Public API

Prefer zero ops? Point at the hosted PelletPilot cloud and skip the setup.

What you get

A grill that answers to you

Live telemetry

Chamber & meat-probe temps, setpoint, fan/igniter/auger activity — decoded app-accurate.

Control

Set target temp, light, prime, power — over the same protocol the vendor cloud uses.

Probe graphs

Time-series meat-probe charts with stall-aware time-to-target estimates.

Open protocol

Fully documented serial + cloud protocol. Build your own client on @pelletpilot/protocol.

Smarter holds

Optional supervisory control for tighter, steadier temperatures.

No lock-in

MIT-licensed. Your data, your server, your grill.

Quick start

Talk to your grill in one command

export PELLETPILOT_GRILL_IP=<your-grill-ip>
python3 tools/pbclient.py state     # decoded live state
python3 tools/pbclient.py watch     # poll every 5s

TypeScript packages (@pelletpilot/protocol, server, app) are in the monorepo. Firmware lives in PelletPilot/firmware.

Status

Roadmap

✔ Reverse-engineered device, serial & cloud protocols (docs 01–08)
✔ Reference client + validated decoder (Python)
✔ @pelletpilot/protocol — TypeScript codec + local RPC client
◻ Server — self-hostable relay + public API
◻ App — mobile PWA dashboard
◻ Firmware — drop-in ESP32 app that repoints your grill to your cloud