Phantom Chat

Phantom Chat is the messenger I built because every other one felt like a compromise. Signal wants my phone number. WhatsApp wants it too. iMessage is tied to my Apple ID. Telegram's encryption is opt-in. Wire wants an email. Threema is decent but the local export still leaks structure. I wanted to message a friend without handing over a permanent identifier first. So I started building Phantom Chat — solo, in evenings and weekends, from Western Australia. Two design choices were locked in from day one: 1. Anonymous sign-up. No phone, no email, no real-name field. You pick a username, that's it. Contact discovery is invite-code-only — there's no way to search for or browse other users. If your invite code never leaves the people you sent it to, your account essentially doesn't exist outside that circle. 2. Post-quantum encryption from day one. Apple shipped PQ3 on iMessage in 2024. Signal added PQXDH the same year. WhatsApp still doesn't have it. I figured if the giants were finally doing it, an indie shipping in 2026 had no excuse not to. Every conversation runs through a hybrid Signal Protocol Double Ratchet + NIST FIPS 203 ML-KEM-768 (Kyber). "Store now, decrypt later" attacks don't work on it. The business model is the part I'm most opinionated about: no VC, no ads, no data sales. I've capped lifetime "Founder" seats at 500 specifically so I never have enough runway pressure to take outside investment or sell user data. Subscribers fund the company. If I run out of subscribers I run out of company — but I never become a different company that wants to sell your messages to a model trainer. The fight with myself for the last six months has been honesty in marketing — I do code-vs-claim audits before every release and have removed three site claims that didn't quite hold up (Sealed Sender, an outdated iMessage comparison, and the word "air-gapped"). I'd rather lose a hook than ship spin. Happy to chat about any of it.