PolyCast
See PolyCast running.
A live demonstration environment, seeded with sample records. It is not your data, and nothing you do in it is kept.
Open the PolyCast demoWhat you’ll see
The parts worth your ten minutes.
One live source becoming several language streams at once.
The delay on each language stream, updating while the session runs.
Interpreter-assist: a person correcting the machine's translation live.
Streams going out to several destinations at the same time.
Access
Access details are issued on request. Ask, and you will get them from the engineer who would run your engagement.
A route through it
If you only have ten minutes.
You do not have to follow this. It is the path that shows the engineering rather than the interface.
Add a source
Start with the microphone source. This is the whole input.
Add target languages
Add three. Each becomes its own stream, running independently — not one output with options.
Watch the latency readouts
Each stream's delay, live. Add a fourth language and watch what does not happen to the other three.
Route to destinations
Send two languages to different endpoints.
Turn on interpreter-assist
See where the person fits in — correcting the machine, not replacing it.
We build systems like this for other organizations.
PolyCast exists because we met the same operational failure enough times to engineer it properly. Yours is probably a different failure. That is the conversation we want, and you will have it with an engineer.