It runs on the rugged handset, not just the phone

Why this hardware suits us

A PoC radio is an Android device with the parts a phone leaves out, and the way they are usually deployed happens to match how push.tt is built.

• No Google Play Services required anywhere in the app — a lot of this hardware ships without them, and a lot of it is deliberately kept off Google's network
• A foreground microphone service, so a transmission survives the screen going off in a pocket
• Sideload and MDM friendly. These fleets are provisioned by an administrator, not from a store — see device management
• End-to-end encryption, which the incumbent in this market does not offer at all. On hardware bought specifically because the work is sensitive, that is the whole argument

The PTT paddle has no standard

This is the part that quietly defeats software written for phones. There is no common way for a rugged handset to tell an app that the big orange button was pressed — vendors do it three different ways, and the codes are not reliably discoverable.

• A key event — works only while the app is in the foreground, which is not how a radio is used
• A broadcast — works with the screen off, and is what most of this hardware actually sends
• A media button — which never reaches an app that only overrides key handling, because the platform routes it to whatever holds the active media session. An app that misses this looks dead to every wired speaker mic
• All three are handled, and the built-in table leads with the pair that PoC firmware ships for the incumbent's sake — the closest thing this market has to a convention
• Plus a learn mode. Press your radio's paddle in Settings and it binds whatever that radio actually sends. PTT and the SOS key are learned separately, because they are separate buttons with separate consequences

Verified on a device: a broadcast paddle driving a 4.3-second encrypted transmission, the app choosing the right conversation with nothing on screen (open conversation, then default channel, then the only channel, then most recent), and a held paddle not firing repeatedly.

The cost you should know about before you buy

One real limitation, stated plainly because it affects your data bill rather than ours.

• Opus encoding needs Android 10. The decoder works from Android 5, but the encoder does not — and PoC radios typically ship Android 8 to 11
• So an older handset transmits PCM16 at about 32 kB/s — roughly nine times the data of Opus — on exactly the metered cellular connection the codec exists to protect. It still receives Opus from everyone else
• Fixing it properly means bundling the codec ourselves rather than using the platform's. That is real work and it is not done
• Ask us what your specific model reports before committing to a fleet. It is one number and it decides your monthly data

Written but not yet proven on hardware

Listed rather than implied, because "supported" and "we wrote the code" are different claims and only one of them is honest here.

• Bluetooth and wired speaker-mic buttons — the media-session path is implemented and has not been confirmed on a real accessory
• Restart after a power blip — a radio in a vehicle cradle should come back by itself; the receiver exists and is unverified
• Small screens. The talk button was sized from screen width alone, which on a 240×320 display made it taller than the space available. It now sizes from the shorter axis, and that has not been seen on the actual screen

These are unverified because the test device degraded rather than because they were skipped. If you have hardware in hand, that is the fastest way to close this list — and we would rather do it with you than claim it in a datasheet.