US Patent 11,315,588 · Active & granted

Turn any learner into
a native speaker

SpeechMatch is a patented real-time speech-coaching engine built to live inside your language-learning platform — and built to transform it.

USPTO-granted patent Protected through 2038
$65B
Global language-learning market today
1.5B
Active language learners worldwide
$300B
Projected market size by 2032
How it works

The world's most precise
speech-coaching engine

SpeechMatch listens the way no human teacher can — analyzing pitch, rhythm, and amplitude in real time, then giving learners instant visual feedback so they can hear and see exactly where they're landing.

Real-time pitch analysis
SpeechMatch captures the pitch-frequency envelope of a native phrase and instantly compares it to the learner's attempt — measuring every inflection point with Fast Fourier Transform processing.
Rhythm & cadence matching
Language isn't just words — it's music. SpeechMatch scores the rhythm and pacing that separate native speakers from foreign-accented ones, tracked independently from pitch and volume.
Amplitude & volume
The system tracks the natural volume dynamics of speech — the rises and falls of amplitude that give language its character and make it convincing to native ears.
Three-strand visual feedback
Learners see pitch, rhythm, and volume as three independent visual strands at once — making the invisible mechanics of speech visible, and learnable, for the first time.
Scored like a game
Each attempt generates individual scores for volume, rhythm, and pitch, plus a composite overall score. Learners watch their numbers climb with every rep — engaging by design.
Seamless SDK integration
SpeechMatch drops into your existing app architecture — smartphone, tablet, laptop, desktop, and cloud. Your interface, your brand; our engine underneath.
US Patent No. 11,315,588 B2
Communication System for Processing Audio Input with Visual Display
The patent covers the full SpeechMatch methodology: processing spoken reference phrases into quantized representations of amplitude, rhythm, and pitch frequency; generating simultaneous visual displays of all three components; receiving user speech; comparing it in real time; and calculating composite match scores. The patent is active, fully granted, and publicly registered with the USPTO.

Platform developers who build similar functionality independently do so at their own legal risk. The smarter path is a licensing partnership — and we're inviting a select group of platform leaders to be first.
Patent number
US 11,315,588 B2
Status
Active — granted
Grant date
April 26, 2022
Anticipated expiration
October 10, 2038
Inventors
Robert Taub & Lawrence Welkowitz
Filed
October 10, 2018
View full patent on Google Patents
The team

Built by the engineers
who created Guitar Hero

SpeechMatch wasn't built by academics or edtech startups. It was built by a team that already cracked the hardest problem in consumer tech: making precision skill-building feel like play.

From Guitar Hero to SpeechMatch
The same feedback-loop genius, applied to speech
The engineering team behind Activision's Guitar Hero turned real musical technique into an addictive, score-based experience — and applied the same core insight to language learning. If you can make a guitarist obsess over hitting the perfect note, you can make a language learner obsess over hitting the perfect pitch. SpeechMatch is the result.
$2B
Guitar Hero global revenue
25M+
Copies sold worldwide
2
Breakthroughs, one team
Inventors

Robert Taub & Lawrence Welkowitz

The patent is held by its inventors, Robert Taub and Lawrence Welkowitz. Their work on SpeechMatch grew from years of clinical research into speech-pattern recognition, autism communication therapy, and the science of vocal congruence — the finding that people who best match the rhythm, volume, and pitch of others are judged as superior communicators.

Applications

Beyond language learning

SpeechMatch has been tested to help children and adults with autism capture the cadence and rhythm of standard speech — supporting them in broader social and professional contexts.

Language & accent learning
The primary application: helping learners of any language sound natural, not foreign. Works for any language pair — English, Mandarin, Spanish, French, and beyond.
1.5B
Global language learners
Autism speech-therapy support
A clinically researched application helping individuals with ASD understand and replicate the volume, rhythm, and pitch patterns of neurotypical speech — opening new doors for communication and connection.
8M+
Potential reachable users globally
The opportunity

A $300 billion market
looking for this technology

Language learning is one of the fastest-growing categories in consumer tech. Every major platform is competing on engagement and outcomes. SpeechMatch delivers both — and the patent means there's only one place to get it.

$65B
Current global market size
1.5B
Active learners worldwide
$300B
Projected market by 2032
2038
Patent protection through
Why license now

First movers win

SpeechMatch technology is going to be inside every major digital language-learning platform. The question isn't whether your users will expect this capability — it's whether your platform will have it first, and have it licensed cleanly.

US Patent No. 11,315,588 is active, granted, and publicly registered. Any platform that independently builds similar functionality does so at legal risk. We're inviting a select group of platform leaders to move first, on favorable terms — as partners, not defendants.

Licensing inquiries

Let's talk about
what this looks like for you

We're looking for a small number of platform partners to move first. If you're building in language learning, this is a 20-minute conversation worth having.

One business day. We read and respond to every inquiry personally.

No pressure. We want to understand your platform before talking structure.

Direct to the inventors. You're talking to the people who hold the patent.

We respond within one business day. Your details are sent securely and never shared.