Why we replaced the NFC tag with a Bluetooth key

Spark started with an NFC tag. You tapped it to your phone, the phone read a token, the apps unlocked. It was pleasingly physical. It also did not work on a large number of phones.

Availability was the forcing function

A large share of budget and mid-range Android phones ship with no NFC radio at all, and on the ones that have it the API surface is uneven, inconsistently exposed to apps, and on some platforms restricted in ways that make a reliable tap-to-unlock flow impossible. That matters more to us than it would to a company selling elsewhere: our first market is India, where mid-range Android is the phone most people are actually holding. A focus tool that only works on some phones is not a product. Bluetooth Low Energy is available essentially everywhere.

What we lost, and had to rebuild

NFC's underrated virtue was not cryptographic — it was practical. Cloning a tag requires hardware and intent. Radio is different: anything broadcasting a fixed value can be copied by software in seconds. Moving to BLE with a naive static token would have been an availability upgrade and a security downgrade at the same time.

So the move came with a protocol change rather than a straight port. The static token became a challenge–response exchange, which restores the guarantee in software instead of relying on the physics of a short-range radio.

• Works on effectively every modern phone, not a subset.
• Range becomes a feature: the key can live in another room.
• New failure mode to design around — the key now has a battery.

The battery question

A dead key must never mean a permanently bricked phone. That is why emergency unlocks exist and are deliberately limited rather than unlimited: enough to handle a flat battery or a lost key, few enough that they cannot quietly become the normal way you unlock.