bothX

bothX is an acoustic media-provenance and authentication technology developed by the Delaware-registered company both WORKS Inc. (also styled both Inc.). It embeds an inaudible signal carrying an identifier, time, date and location data in speech, music or other audio, so that an authorised verifier can later check where and when the audio was produced. The method, which the company calls "Hybrid Spectro-Morse" acoustic authentication, is the subject of United States Patent 12,626,708 B1, granted on 12 May 2026.

Technology

Signal design

According to the patent, bothX encodes a rolling acoustic signature across audible and inaudible frequencies, from 0 Hz to 30 kHz and "scalable to 124 kHz". Each signature carries an identity code together with the time, date and GPS location. The signature is modulated using one or more of amplitude modulation, frequency modulation, phase modulation, frequency-shift keying (FSK), phase-shift keying (PSK) and a Morse-style pulse-duration modulation, combined with time-division multiplexing. The "Spectro-Morse" name refers to this pairing of frequency-domain keying with Morse-like pulse lengths. The patent describes payload characters mapped directly to tones or pulse durations rather than to a hash. It also describes concealing the signal with acoustic dithering, noise shaping and randomised scheduling, and an "anti-jamming" mode that hops the signature between sub-bands.

Decoding, as described in the patent, uses band-pass filtering and time-frequency analysis, such as the fast Fourier transform, the short-time Fourier transform or a bank of Goertzel filters tuned to the expected carriers, followed by checksum or error-detection validation.

Custody and key management

The patent describes a "dual custody" architecture: each authenticated recording is split into complementary, bit-interleaved bit-interleaved segments, with one stored on the user's device and the other in a cloud database, so that neither party alone can decrypt it. In one claimed form, the decryption key is divided into five shares using a 3-of-5 Shamir's secret sharing scheme, and each decryption event is recorded on an immutable ledger. The specification also describes a fallback mode that emits a reduced signature, containing only identity and timestamp, when the cloud segment is unreachable.

Implementation

The patent describes software implementations for mobile, web and embedded platforms that work through conventional speakers and microphones. It also describes a dedicated emitter: a coin-sized, NFC-provisioned piezoelectric device tuned to 30 kHz or above, with an on-board controller for timing. The patent states that this transducer can produce inaudible carriers with a higher signal-to-noise ratio than typical phone speakers. Several claims and embodiments specify that marked audio be kept as uncompressed WAV at a sampling rate of 48 kHz or higher, and the specification gives a minimum signal-to-noise ratio of 20 dB for reliable decoding.

The company's website describes an optional consumer device "about the size of a dollar coin", carried as a pendant, keychain or pocket item. According to the company, it pairs over Bluetooth with a bothX identity wallet app, uses a high-frequency speaker to emit rolling bothX provenance marks that a compatible verifier can recover, and adds on-device GPS location context when available. The company states that acoustic reception and GPS availability depend on operating conditions, that "recovering a mark alone does not authenticate a person or location", and that "the existing evidence covers recovery of encoded data from controlled, software-generated audio."

The company describes the mark as one that cannot be edited without audibly degrading the file, and as verifiable offline. These performance claims are the company's own; no independent evaluation has been published.

Patent

both Inc. of Middletown, Delaware, filed U.S. application No. 19/314,667 on 29 August 2025. It was granted as U.S. Patent 12,626,708 B1, "Hybrid spectro-morse acoustic authentication system and method", on 12 May 2026, with classifications G10L 19/018 and H04L 9/08 and 9/32. The named inventors are Mulakkal Thomas George, Bosky Cherin Varghese and Senthil Kumar Muthian, all of Dubai, and the patent is assigned to both Inc. Google Patents lists the patent as active, with an anticipated expiration date of 29 August 2045. An earlier international application, PCT/IB2023/059224, was cited in the company's 2024 announcement. The U.S. patent lists its own filing date as its priority date.

The patent contains 15 claims, three of them independent:

• Claim 1 (method): emitting an inaudible signature containing a unique identifier, time, date and location coordinates; modulating it with one or more of the listed schemes; overlaying it on audio content as one or more carriers in a designated band, with the identifier, time, date or location encoded as alphanumeric symbols mapped directly to tones or pulse durations; splitting the encoded audio into complementary segments stored on a device and in a cloud database; decrypting them with a key; and recording each decryption event on an immutable ledger.
• Claim 7 (system): an NFC-enabled piezoelectric emitter that plays the signature, together with the dual-custody, shared-key storage architecture, where the audio is verified by checking the embedded identity, time, date and location.
• Claim 12 (method): as in claim 1, but with the signature overlaid as a single carrier in the ultrasonic band, the key split under a 3-of-5 Shamir scheme, and a specified decoding and verification process.

Dependent claims add sub-band shifting across 0–124 kHz, an optional date-of-birth field, uncompressed WAV storage at 48 kHz or above, concealment by dithering and noise shaping, embedded parity blocks, and Goertzel-filter decoding.

The specification acknowledges earlier work on audio watermarking and authentication, including a 2024 published application by Alattar and others that combines media watermarks with blockchain records against deepfakes. It states that none of the prior art discloses "a lightweight, scalable system and method of encoding a legible digital identity within high frequency audio content" using the listed modulation schemes with time-domain multiplexing.

Standards and interoperability

both WORKS Inc. is listed as a member of the Coalition for Content Provenance and Authenticity (C2PA), which the company says it joined as a Contributor Member in 2026. The C2PA conformance explorer lists "bothX C2PA Evidence Bridge" from both WORKS Inc. as a conformant validator product, with a conformance date of 17 August 2026.

In September 2025 the company applied to the MP4 Registration Authority to register an ISO base media file format brand for a "Static Acoustic Verified" profile. The profile is a file that pairs still visual content with a timed acoustic-verification track. The four-character brand SAV1, referencing the company's SAVX specification, appears in the authority's register of brands.

Context

Digital watermarking embeds information such as origin, ownership, timestamps or unique identifiers in the content itself. Provenance metadata instead carries that information in a separate channel, which can be stripped or altered. Audio watermarking has a long history in broadcasting. More recent systems such as Meta's AudioSeal use neural networks to embed imperceptible marks in speech and detect AI-generated audio.

The C2PA's Content Credentials attach a cryptographically signed manifest describing an asset's origin and edits. The specification allows a "soft binding", such as an invisible watermark, to recover the manifest if the metadata has been removed. The C2PA notes that Content Credentials show whether provenance data is intact and signed by a trusted party, not whether the content is true.

Research on near-ultrasonic signalling between consumer devices has found that ordinary speakers and microphones are designed for the audible range. It also found that lossy compression such as MP3 or AAC at common bit rates, and some broadcast chains, can remove frequencies above about 18 kHz.

Development

bothX was developed by both, a company co-founded by M Thomas George and Bosky Cherin Varghese. The company name combines the first letters of the founders' given names ("BOsky" and "THomas"). George, who describes himself as the company's technology director and is the first-named inventor, previously worked in IT audit and cyber security. Cherin, the company's creative director, previously worked in advertising. Senthil Kumar Muthian is the third named inventor. The company also operates Vibrue, a sound-based wellness brand.

The technology was announced in a September 2024 press release from both FZ LLC. The release described it as patent-pending and said pilots were under way in several industries. In September 2026 the funding database VCBacked reported a US$500,000 seed round for the company.

External links

• Official website
• US 12,626,708 B1 at Google Patents

This article "BothX" is from Wikipedia. The list of its authors can be seen in its historical and/or the page Edithistory:BothX. Articles copied from Draft Namespace on Wikipedia could be seen on the Draft Namespace of Wikipedia and not main one.

---