Pond unifies every byte your household generates — smart bulbs, thermostats, routers, solar panels, cameras — into a single queryable lake. SQL in. Insight out.
Built on columnar storage and embedded DuckDB. No cloud required — runs on the Raspberry Pi in your cupboard.
Join temperature logs with your grocery deliveries. Ask your house questions like it's a database — because now it is.
Zigbee, Z-Wave, Matter, MQTT, Home Assistant, Prometheus — 80+ first-party connectors and a 12-line SDK for everything else.
Parquet under the hood. Back up to any S3 bucket, or keep everything on your NAS. Your data never leaves the house unless you say so.
Vectorized execution, pushdown predicates, and column pruning. Scan 10 years of sensor data on a Pi 5 in under 400ms.
Run pond up and it auto-discovers every compatible device on your network. No YAML, no daemons, no drama.
MIT licensed. No telemetry. No account required. Self-host in 60 seconds or deploy the managed edge runtime to 100+ sites.
Point Pond at your devices, open a SQL shell, and start asking real questions about real life.
-- when is the living room coldest this week? SELECT date_trunc('hour', ts) AS hour, avg(temp_c) AS avg_temp, count(*) AS samples FROM sensors.thermostat WHERE room = 'living_room' AND ts > now() - INTERVAL '7 days' GROUP BY hour ORDER BY avg_temp ASC LIMIT 5;
| hour | avg_temp | samples |
|---|---|---|
| Tue 05:00 | 16.2 °C | 60 |
| Wed 04:00 | 16.4 °C | 60 |
| Mon 06:00 | 16.7 °C | 60 |
| Thu 05:00 | 16.9 °C | 60 |
| Fri 04:00 | 17.1 °C | 60 |
A single binary. Runs on macOS, Linux, and every Raspberry Pi since the 3B+.
$ brew install pond
Auto-discovers everything on your LAN. Add anything exotic with a one-line source config.
$ pond add --auto
Open the shell, point Grafana at it, or hit the HTTP API. Your data, your way.
$ pond query
Free forever for home use. No signup, no telemetry, no nonsense.