proflame

The SIT Proflame 2 fireplace remote protocol, in Python: encode a complete appliance state into OOK timings, decode received timings back into a state and the identity of the handset that sent them, and derive the per-remote checksum constants from a single captured frame.

Transceiver-agnostic by design: this library deals in signed-microsecond timing lists, and a radio is whatever turns those into RF and back. The first consumers are the hass-proflame Home Assistant integration and the hackrf-proxy daemon ecosystem, but nothing here depends on either.

from proflame import Remote, State, encode_timings, decode_frame

# One captured frame is enough to learn a handset, checksum constants and all.
heard = decode_frame(received_timings)
remote = heard.remote

# A frame always carries the complete appliance state, never a delta.
timings = encode_timings(remote, State(power=True, flame=4, light=2))

Two things worth knowing before building on it:

• The checksum constants are per handset. They fall out of any one valid frame (Remote.from_frame, derive_key); hardcoding one remote's constants produces frames every other receiver rejects.
• The appliance is stateless and answers no questions. Every frame carries the whole state, so a consumer must hold its own belief; nothing in the protocol confirms what the appliance did.

Documentation and data

• The protocol derivation — physical layer, framing, field layout, checksum model, open questions, acknowledgments — lives in docs/PROTOCOL.md.
• The field-confirmation procedure and its results live in docs/MAPPING.md.
• Real captures — a community five-remote packet table (aggregated from rtl_433#1905) and frames recorded off the air, used as golden test data — live in tests/.
• Bench scripts — the standalone reference decoder and the checksum analysis that first solved the relation — live in tools/.
• Related work — smartfire, rtl_433's decoder, HACS-Proflame2 — is compared and credited at the end of docs/PROTOCOL.md.

License

MIT OR Apache-2.0, at your option.