The problem
Carriers buying wholesale voice routes had little hard evidence of how those routes performed.
A degraded route, a vendor that answers early and bills for silence, or a call that connects with no audio often surfaced first as a customer complaint. Checking a route meant placing calls by hand and reading switch logs, with nothing structured to send back to the vendor.
What we built
ThinkTank AI built Nurlinks a private platform that places real test calls and scores every one.
The Nurlinks team adds its SIP trunks and a list of destination numbers, and Zboun dials them from its own Asterisk server within each trunk's channel and call-rate limits. A Node.js service captures packets and audio during each call, measures RTP, silence and ringback, and estimates MOS. The Laravel app builds a report that updates live as calls finish.