Introducing Polythings
What we're building, who it's for, and why it's built on open source we maintain.
Every connected product needs the same things under it: a way to give each device an identity, update it safely, collect its data, keep that data, make sense of it, and tell someone when something goes wrong. Most teams build these themselves, one service at a time, and then spend years running them. Polythings is that foundation, as one platform.
What we're building
Building blocks behind one API, each usable on its own:
- Devices and a certificate authority: identity for every device, issued at the factory and renewed automatically.
- Over-the-air updates for microcontrollers, embedded Linux, devices behind gateways, and ML models, rolled out in waves that stop themselves when something looks wrong.
- Ingestion and storage for MQTT, HTTP, CoAP, LwM2M and industrial protocols, kept as open tables in our bucket or yours.
- Assets, metrics and alarms: model the machines your devices measure, define what every number means, and alarm on it with a proper lifecycle.
- Analytics, serverless jobs and predictive maintenance: SQL, dashboards, your own code next to the data, and health scores for every machine.
- Cellular connectivity and migration: SIMs managed next to devices, and a way off the platforms teams outgrow.
The idea comes from WorkOS, which turned enterprise features into small, excellent APIs. We want the hard parts of IoT to feel the same: an afternoon's work, not a quarter's.
Who it's for
Companies that build equipment and sell it into plants and buildings: pumps, compressors, HVAC, process and packaging machinery. They are connected-product companies and industrial companies at once. They want to sell uptime and service, they ship devices that need updates, and from December 2027 the Cyber Resilience Act makes that last part a legal duty in the EU.
Why open source underneath
Everything that runs on your devices, and every format your data is stored in, is open source under Apache 2.0. The storage layer, timelake, is already public. The commitment is simple: you can always take your data and your devices with you.
Where we are
In design. The plan for the platform, the roadmap and the benchmarks are public, and we'll write here as we build. If you're running a fleet on Azure IoT Hub, AWS IoT Core, Particle or a platform you built yourself, the migration guide shows how the move will work.