Cell Towers Are Optional: At 16, I Built a Hybrid 800m Off-Grid Mesh Protocol for Smartphones (No LoRa Required)

How I combined BLE, Wi-Fi Direct, UWB, and Delay-Tolerant Networking into AirPin — a zero-server P2P engine that turns regular phones into an encrypted relay network.

What happens when cell towers go down, Wi-Fi is dead, or you are deep in the wilderness—and you don't have specialized LoRa hardware like a Meshtastic node in your backpack?

Every modern smartphone already carries a ridiculous amount of radio silicon: Bluetooth Low Energy (BLE), Wi-Fi Direct, and Ultra-Wideband (UWB). Yet almost every off-grid messenger on the market locks itself into a single transport layer (usually BLE), suffering from terrible bandwidth, short range, or rapid battery drain.

Over the past year, I decided to fix this. I designed and built AirPin—a hybrid, multi-protocol P2P mesh routing engine that dynamically switches between radio interfaces and uses Delay-Tolerant Networking (DTN) to bounce encrypted messages from phone to phone without the internet or central servers. It serves as the core engine of my upcoming super-app ecosystem, Prux.

Here is the story and technical breakdown of how AirPin works under the hood, how multi-hop relay behaves in real-world field tests, and the architectural problems I had to solve along the way.

Why Single-Radio Offline Messengers Fail

If you look at classic off-grid apps like Bridgefy or Briar, they hit a hard physics wall:

• BLE-only mesh: Great for background discovery and low battery usage, but throughput is tiny, and real-world range drops sharply in dense environments.
• Wi-Fi Direct: High bandwidth and up to 200+ meters line-of-sight range, but keeping Wi-Fi Direct awake 24/7 destroys smartphone batteries and triggers aggressive OS background killers on iOS and Android.

Instead of choosing one radio, I built AirPin as a dynamic link-layer router that treats BLE, UWB, and Wi-Fi Direct as a single adaptive pipeline.

Inside the AirPin 4-Layer Stack

AirPin operates automatically in the background across four distinct stages, choosing the right radio for the right moment:

1. Always-On Discovery: Devices running Prux broadcast lightweight, rotating cryptographic beacons over Bluetooth Low Energy. Because BLE uses almost zero power, it acts as a silent radar. Short text messages and initial handshakes go straight through BLE without waking heavier radios.
2. Spatial Ranging & Direction (UWB): When supported by the phone's hardware, Ultra-Wideband gives centimeter-accurate distance and directional vectors between peers. Instead of blindly flooding packets to every phone nearby, AirPin knows which nodes are physically moving toward the destination or holding the strongest link.
3. High-Speed Burst Escalation (Wi-Fi Direct): When you send voice notes, images, or need maximum single-hop distance (up to 200+ meters in open terrain), AirPin negotiates a handshake over BLE, spins up a temporary Wi-Fi Direct P2P tunnel for a few seconds, blasts the encrypted payload across, and immediately puts the high-power radio back to sleep.
4. Store-and-Forward Relay (DTN): What if there is no continuous chain of users between you and your friend right now? AirPin uses Delay-Tolerant Networking. A message hops across intermediate phones like a baton in a relay race (reaching 750–800 meters across 4 hops in field tests). If the chain breaks, an intermediate phone securely caches the encrypted packet in an isolated sandbox and forwards it minutes later when it walks past another user.

Zero-Server Security: Can Strangers Read My Relayed Messages?

Bouncing your private messages through strangers' phones sounds terrifying at first. To make it bulletproof, AirPin enforces strict cryptographic isolation:

• End-to-End Encryption (E2EE) at the Edge: Messages are encrypted on the sender's device before they ever touch the AirPin router. Intermediate relay phones only see an opaque encrypted envelope, a randomized packet hash (to prevent infinite broadcast loops), and a hop counter (TTL).
• Zero Cloud Storage: Off-grid messages never touch a central server. Even when Prux operates in standard online mode, message history stays strictly on user devices.
• SunFlower Visual Codes: You can't trust a central key server when you are offline. I created custom visual SunFlower codes (a specialized alternative to standard QR codes) so users can verify identities and exchange encryption keys face-to-face in a split second.

Real-World Field Tests & Recognition from Chipmakers

Testing a mesh protocol in an IDE simulator is easy; testing it in the real world is brutal. I live on the Pacific coast of the Russian Far East, surrounded by steep hills, dense taiga, and coastline where dead cellular zones are an everyday reality.

In field tests across urban and open terrain, AirPin achieved:

• 40–80 meters single-hop range indoors and in dense city blocks.
• 200+ meters single-hop range in open line-of-sight.
• 750–800 meters end-to-end delivery across a 4-hop chain of smartphones without internet or cell service.

Earlier this year, I pitched and defended the AirPin architecture in English at the 40th CASTIC international science and engineering competition in China. Out of 600+ projects, AirPin placed in the Top 20 and won the Unisoc Innovation Award directly from UNISOC, one of the world's largest mobile processor and baseband chipmakers. The technology is also officially registered with Rospatent (Software Reg. #2026607225, Patent Pending #2026112785).

Why I Wrapped a Mesh Protocol Inside a "Liquid Glass" Super-App

Here is the dirty secret of every standalone offline messenger: nobody has it installed when an emergency actually happens. People delete single-purpose emergency apps because they don't use them on a normal Tuesday.

To make a smartphone mesh network work in the real world, you need node density—people actually keeping the app on their phones every day. That is why I built Prux around AirPin, featuring a custom "Liquid Glass" UI for iOS, Android, and macOS. Alongside the off-grid messenger, Prux includes daily-driver mini-apps:

• Prux Mail: Built-in email client aggregating multiple accounts in one clean inbox.
• Loud Music: Offline-capable music player.

• ZheykPay: Integrated crypto wallet.

• Red Panda & Mahjong Blast: Language learning tools and mini-games.

The more people use Prux for music, mail, or everyday chatting, the stronger and wider the invisible AirPin mesh network becomes whenever the grid goes down.

Follow the Build & Join the Beta

I am currently running closed beta tests and fine-tuning background BLE/Wi-Fi Direct wake-locks ahead of the public launch in Spring 2027.

I'm building Prux and AirPin in public and sharing raw engineering logs, range test videos, and UI previews over on X. If you love mesh networking, P2P protocols, or mobile systems engineering, come say hi and follow the journey:

• Follow me on X (Twitter) for live dev updates: https://x.com/Prux_Messenger
• Telegram Channel: https://t.me/Prux_channel
• Official Website: https://www.prux.su

Questions for the devs here:

• If you've fought iOS and Android background execution limits for BLE/Wi-Fi Direct, what are your cleanest workarounds for keeping discovery alive without killing the battery?
• Would you be interested in an open standalone SDK for AirPin so you could drop offline mesh messaging into your own mobile apps?