Legal
Security
How the system is built, where data travels, and what that means for a confidential event.
Processing happens in your building
Speech recognition and translation run on hardware we bring into the
venue. There is no cloud API in the path, no audio upload, and no vendor
other than us with access to what is said. Once set up, the system works with
the venue's internet disconnected.
For government briefings, legal proceedings, medical conferences and
commercially sensitive meetings, this is usually the deciding factor.
The relay, and what it carries
Guest WiFi frequently blocks devices from reaching each other, and many
attendees stay on cellular. When that is the case, caption text
is published through our server so those phones can read along.
- It carries text only. Audio never leaves the venue.
- It is off by default and enabled per event, by the
operator, with your agreement.
- It stores no captions. Only attendance counts and any feedback given
are written down.
- Traffic is encrypted in transit.
If your event cannot have caption text leave the building at all, we run
without the relay. Attendees then read on the venue network or on the
screens, and nothing crosses the boundary.
Event links
An attendee link contains a 160-bit random token. It is not guessable and
not indexed by search engines. Anyone holding the link can read that event's
captions, which is the same trust model as a printed code handed out at the
door. Worth knowing if your event is closed.
Our equipment
Machines used at events are ours, kept up to date, and used for nothing
else. Transcripts are removed after hand-over as described in the
privacy notice.
Reporting a problem
If you believe you have found a security issue, tell us at
speakmany.com/contact and we will respond within one
working day. Please give us a reasonable chance to fix it before publishing.