How connect-by-code remote access works (no accounts, peer-to-peer, encrypted)

MetaRemote connects two machines with nothing more than a 9-digit code and a password. Here's what actually happens under the hood — and why the design keeps your session private.

1. The host gets a stable code

When you install the host on a machine, it registers with the relay and is assigned a stable, random 9-digit code (shown in its menu-bar panel). The same machine keeps the same code across restarts, so you can save it like a phone number.

2. The viewer routes by code

On the viewer you type the code. The relay uses it purely to route — to introduce the two machines to each other. Viewers don't need an account or a pre-shared token; the code is the address.

3. Password is checked peer-to-peer

The unattended password you set on the host is never sent to the relay. Once the two machines connect, the viewer sends the password over the already-encrypted channel and the host checks it against a stored argon2 hash. The relay is zero-knowledge about your password.

4. Media goes direct, encrypted

MetaRemote gathers connection candidates and prefers a direct peer-to-peer path (falling back to a relayed path only if a direct one can't be established), then streams hardware-encoded HEVC video and input over DTLS-SRTP. The relay carries signaling, not your screen.

5. Optional "ask before allowing"

Hosts can require a person to approve each incoming session, on top of the password — useful for attended support where someone is sitting at the machine.

That's the whole model: a code to find the machine, a password verified directly between the two ends, and an encrypted peer-to-peer stream. New to MetaRemote? Start with what MetaRemote is, or request early access.