NoChat is engineered on a zero-trust model: the network, the server, and even NoChat itself are all treated as untrusted. Every message is encrypted on-device and stays that way.
Security claims are easy to make and hard to verify. NoChat is a secure messaging app where the guarantees are enforced at the protocol level, not promised in marketing copy. The server stores encrypted blobs and routes them — that's its entire job. Keys are generated on your device, content is end-to-end encrypted, the cryptography is standardized and auditable, and the whole codebase is open source so you can check every claim yourself.
A secure messaging app is only as strong as its server architecture. If the service holds your private keys, your encryption is theater. NoChat's server stores encrypted ciphertext, routes it, and nothing else. Identity keys, exchange keys, and session keys are all generated on your device — the server never sees them and could not derive them if it tried.
Being specific about your threat model is the difference between a serious secure messaging app and a marketing one. Here's what NoChat protects against — and the one thing no app can.
NoChat uses primitives anyone can audit: AES-256-GCM for content, P-256 ECDH for key agreement, HKDF-SHA256 for key derivation, and P-256 ECDSA for signatures. Calls use WebRTC with DTLS-SRTP. Post-quantum key encapsulation (ML-KEM) is implemented on the backend and on the roadmap for the client — we describe it as planned, not deployed, on the frontend.
Every line of the client and server is public at github.com/kindlyrobotics/nochat. Don't trust our word — read the code, trace where keys are generated, and confirm the server schema only holds ciphertext. The full cryptographic breakdown lives in our crypto inventory.
The service can't read your messages even if it wanted to, metadata is minimized, the cryptography is standardized and auditable, and there's no single point of trust. NoChat is built to hit all four.
Network attackers (ISPs, hotspots, on-path observers), a compromised NoChat server, and tampering in transit. The only thing it can't protect against is a malicious endpoint — if your own device is compromised, no app can save you.
There's nothing useful to steal. The database holds encrypted ciphertext with no decryption keys attached. A full dump gives an attacker exactly what it gives us: opaque blobs.
Under valid legal process we can produce account metadata where it exists, but message content is end-to-end encrypted and we have no ability to decrypt it. That's a mathematical fact about the design, not a policy choice.
The codebase is fully open source and under continuous community review, and the cryptographic primitives are documented in the repository's crypto inventory. Formal third-party audit status is tracked on GitHub.
No. NoChat is free to use and free to self-host. Security shouldn't be a paid upgrade.
Free, open source, and private by design. No phone number, no email required.