“"Walked around Fusion Festival with Bitchat running. Saw about 20 other devices and a few messages. Nobody replied to my messages. So usage isn't zero, but 20 devices out of 80000 people isn't a lot. Most hops I saw was 2, meaning there wasn't enough density to make it any meani...”
You know that feeling when the mobile data goes dark during a protest, a festival, or a natural disaster, and every messaging app you own becomes a blank screen? WhatsApp, Signal, and Telegram all assume a working route to a server, and they all assume you handed over a phone number to get an account. The moment a regulator flips a switch or a tower saturates, that entire stack is gone. bitchat's bet is that the phones around you are the network — your handset becomes both a client and a relay, and messages hop device to device without asking anyone's permission.
Think of it like passing a note across a crowded room — you hand it to whoever is closest, they hand it on, and it reaches the far side without anyone needing the room's PA system. Every phone running bitchat acts as both a Bluetooth central and a peripheral at once, flooding packets outward with a hop limit of 7, so a message can cross a distance no single Bluetooth link could span. Private messages get encrypted with a Noise XX handshake using Curve25519 key agreement, ChaCha20-Poly1305, and SHA-256, so relaying phones carry ciphertext they cannot read. The clever part is that this is not a Bluetooth app: two transports sit behind one `Transport` interface and a `MessageRouter` picks between them, so when the internet is reachable the same encrypted conversation continues over Nostr relays as NIP-17 gift-wrapped events. Since v1.6.0 there is also a store-and-forward layer where a phone can carry sealed mail for someone who is offline, tagged only with a 16-byte rotating recipient tag, capped at 16 KiB and 24 hours.
If you work on offline-first or disaster-resilient communications, or you have ever needed a channel that survives a regional internet shutdown, this is the reference implementation worth reading — the whitepaper and privacy assessment are unusually candid about what the protocol does and does not protect. It is also a genuinely interesting Swift codebase if you want to see a BLE flood mesh and a Nostr client behind one routing abstraction. It is not for you if you need Android or Linux from this repo (Android is a separate GPL-3.0 codebase), if you need a named external security audit before...
Worth exploring and worth installing, but not worth betting anyone's safety on yet. The engineering signals are strong — v1.7.0 shipped 2026-07-08, three releases went out in three days in early July 2026, and the project has SwiftLint and Periphery dead-code checks wired into CI — but the security posture has real gaps: PR #1345 fixed '3 critical, 7 high' audit findings in v1.5.4 without naming an auditor or publishing a report, and PR #1349, a mesh identity-spoofing fix, was still open as of 2026-07-26 with a bot review flagging a bypass in the fix itself. The sibling Android repo under the same org still tells you in its README not to rely on its security until it has been reviewed.
Deep-dive insight, Easy and Pro modes, plus action playbooks — the full breakdown is one tap away.