I had a new smart gas meter fitted and wanted proper gas monitoring in Home Assistant. The first useful step was checking the meter model, the label and the install paperwork before touching SDR gear. That saved a fair bit of time. Start with what the meter actually speaks, not with guesswork about antennas and containers.
Most modern meters in the UK and Europe use a radio interface such as Wireless M‑Bus. It often runs around 868 MHz for smart metering. Recent product notes and RF module vendors back that up, which is why so many DIY SDR setups start there. Read the meter label or ask the installer which protocol and frequency it uses. If the meter lists a vendor protocol, model number or RF spec, that is the clue you want.
If the meter broadcasts on a common RF protocol, try passive capture with rtl_sdr. I have run containers with rtl_sdr, metermon and rtl_433. Start with a proper antenna for the target band and scan across 433 MHz and 868 MHz. Put the dongle near an external wall or window, not in a metal cupboard. Run a long capture while the boiler or cooker is in use, since that makes bursts more likely to show up. If packets appear, metermon or rtl_433 can decode common meter formats. If nothing turns up, change the antenna, move the dongle closer to the meter, then leave it running longer. Battery-powered meters can be slow to talk.
Radio decoding is not always the right route. If the meter exposes a pulse output terminal, a pulse reader is the simplest and most reliable option for gas monitoring. I used a small optocoupler or reed switch on the pulse terminal and fed that into a microcontroller that publishes MQTT to Home Assistant. That gives immediate, high-resolution readings and works well with automation and energy dashboards. Optical reading with a cheap camera aimed at the dial or LED pulse is another option when wiring is not accessible. Some vendors provide an API or a gateway that publishes readings. If that exists for your meter model, it is worth asking the utility. If the meter uses a proprietary vendor cloud, check whether the vendor has an open API or whether the community has already built a Home Assistant integration.
Once you have a data feed, set up the Home Assistant sensors. For MQTT, publish the pulse counts and create a sensor that converts pulses to cubic metres. For decoded RF messages, push the parsed payload into MQTT or the Home Assistant REST API. Keep raw packet dumps while you test so you can reprocess them if the decoder misses fields. For automation, use sensor templates to calculate instantaneous flow and cumulative consumption, then use those for alerts or cost tracking.
Practical takeaways: read the meter label first, then decide whether to use radio capture or a physical pulse feed. If you go the rtl_sdr route, focus on the correct band, use the right antenna and leave it running long enough. Try rtl_433 and metermon before deciding the meter is not speaking. If radio fails, pulse readers, optical readers or vendor APIs usually work and plug cleanly into Home Assistant.
