Video protection
security architecture
seven independent layers stand between your course video and unauthorized redistribution. defeating one layer yields nothing: each assumes the others can fail.
encrypted media at rest
every lesson is stored and delivered as AES-128 encrypted HLS. files on servers and CDN edges are ciphertext; a leaked segment file alone plays nothing.
token-gated key delivery
decryption keys are never public. the key endpoint requires a signed, short-lived token minted only for an authenticated, entitled viewer.
asset-bound tokens
each playback token is cryptographically bound to one specific video. a token issued for lesson A can never fetch the key for lesson B.
signed segment urls
individual video segments are unfetchable without valid, per-session authorization. copying a url out of the player gives an attacker nothing durable.
short-lived credentials
playback authorization expires quickly and is refreshed continuously for legitimate viewers. shared links die fast by design.
session authentication
everything sits behind account-level auth: signed JWTs, refresh rotation, and server-side session controls in front of every media route.
identity watermarking & client hardening
each stream renders the viewer's own identity into the frame, repositioning continuously. screen-capture APIs are intercepted, and automation signals are denied playback.
Honest limits
what this does, and what nothing can do
what encryption stops
downloading a playable file, sharing working stream links, scraping segments, replaying expired sessions, and cross-account key reuse.
what encryption cannot stop
pointing a camera at a screen. no vendor can prevent that, which is why every frame carries the viewer's identity. a leaked recording traces back to the exact account that made it.
deterrence, not theater
client-side defenses (capture interception, automation detection) raise the cost of extraction. we treat them as one layer of seven, never as the security boundary itself.
reporting a vulnerability
found a weakness in any of these layers? we want to know before anyone else does. email datatechsonline@gmail.com with reproduction steps and we'll respond within two business days. good-faith research is welcome; we don't pursue researchers acting responsibly.
questions about protecting your own course content? talk to us.