Auto Secure LoginPlatform
Live Peer-to-peer Bilingual EN / ES

Small-group video where the call itself never touches a server.

ASL Meet is peer-to-peer video for one-on-ones and small groups. A host signs in and gets a short, unguessable room link; guests join with just the link and a display name — no account needed. The media is DTLS-SRTP encrypted directly between browsers, with a relay used only for encrypted packets when the network blocks a direct path. The server stores meeting metadata only — never audio, video, chat, or recordings.

browser-to-browserDTLS-SRTP encrypted media
0call content stored on the server
0npm dependencies
on-devicerecording, saved to your machine

The problem

Most "secure" video still routes your call through their servers.

Mainstream conferencing tools terminate your audio and video on their infrastructure — where it can be recorded, retained, or subpoenaed — and make guests create accounts just to join a call. For a quick, private conversation, that's a lot of exposure.

ASL Meet keeps the media between the participants. The server only helps the browsers find each other; the call content never lands on it. The "nothing stored" claim is architectural, not a policy promise.

Private by construction

  • Media flows browser-to-browser, DTLS-SRTP encrypted
  • Server persists meeting metadata only
  • Recording is client-side, saved straight to your device
  • Everyone sees a recording notice when it's on

Capabilities

Quick, private calls — nothing to install, nothing retained.

🔗

Link-based rooms

A short, unguessable room at /r/<code> — guests join with just the link and a display name, no account.

🔒

Encrypted P2P media

A WebRTC mesh for up to four participants, media encrypted directly between browsers.

🛰

Relay only when needed

Short-lived HMAC TURN credentials are issued only for active meetings — and carry only encrypted packets.

🎛

Host controls

Lock the room, remove a participant, or end the meeting for everyone.

On-device recording

Fully client-side recording composites the call and mixes audio locally, saving straight to your device — with a recording notice shown to all.

🖥

Screen share & chat

Screen sharing plus live in-call text chat that is never persisted anywhere.

📲

Installable PWA

Installs like an app with an offline app shell, plus a native Android build.

Bilingual

English/Spanish with a persistent language selector.

🗑

Metadata that expires

30-day meeting-metadata retention, with active rooms excluded from the sweep.

Node 22built-ins only, zero npm deps
real JWKSfrom-scratch OIDC client, verified
28 testslaunch-readiness, TURN, locale-parity, PWA
fails closedon a weak session secret
Honest about the ceiling. Meet's README states plainly that going beyond four participants would require an SFU that briefly touches media on a server — so it caps at four to keep the server-blind guarantee true. A dependency-free codebase makes that guarantee unusually auditable.

A private call, in one link.

Encrypted peer-to-peer video with no accounts to join and no call content on the server — just share the link.