Fundraising Fox
OfSpectrum
Techstars '25Los Angeles, US · Founded 2024 · 10 employees on LinkedIn · 1 known investors
Find your way into OfSpectrum
Sign up to see every warm intro you have to OfSpectrum
- Paths you didn't know you had: your email and LinkedIn already hold routes to the OfSpectrum team. We find them for you
- 2nd- and 3rd-degree connections: the friend-of-a-friend routes that take hours to manually find through your inbox or LinkedIn
- Answers now, not in days: momentum is everything in a raise. Skip asking around whether someone knows someone
Days of research, done the moment you sign in, with every route ranked by how warm it is.
OfSpectrum builds watermarking and verification technologies to protect the integrity and ownership of digital media content, answering who created it, who owns it, and how it can be used. The company serves record labels, music publishers, streaming platforms, marketplaces, and AI labs with solutions for tracking, authenticating, and preventing unauthorized use of creative work including AI-generated audio.
Also known as OfSpectrum
Founders & leadership
OfSpectrum was founded in 2024 by Yechen Wang and Yishu Huang.
Investors · 1
How we know: the Techstars portfolio · open dataset · Top 3000 Accelerator Startups (no YC) 2026-08 · Not right? Tell us
Company profile
researched Sep 2026OfSpectrum develops watermarking and verification technology for digital media, with a focus on audio. Its core product embeds inaudible data — text, audio, images or database references — directly into an audio signal so that the origin and authenticity of a recording can be verified later. The company describes the watermark as robust to common editing, compression and over-the-air re-recording, and as supporting an information-embedding channel for distributing multimedia data privately. A companion authenticity monitoring service scans the internet for watermarked content to support piracy detection and verification.
Alongside watermarking, OfSpectrum offers deepfake detection that distinguishes AI-generated voices from human voices in real time, including frame-level identification of AI-generated sub-sequences within a recording, and models optimized to run on mobile devices. A provenance tracking layer attaches C2PA-compatible metadata — ownership, creation timestamp, licensing terms and modification history — to the watermark itself, which the company positions as supporting machine-readable labeling of AI-generated and synthetic audio under EU AI Act Article 50 transparency requirements. The company also describes text-to-speech technology built around consent-based voice creation, governed training data and integrated watermarking intended to prevent unauthorized voice cloning.
The company frames its positioning as API infrastructure rather than an academic project or a closed enterprise appliance, citing sub-100ms verification latency, a RESTful API and a Python SDK, identity-based matching that it describes as more efficient than fingerprint comparison, and psychoacoustic embedding validated on ViSQOL and PESQ at 48 kHz.
Business model
OfSpectrum sells access to its watermarking, detection and provenance technology as API infrastructure, with a self-serve developer path (account creation and API key issuance, plus a Python SDK) and a separate enterprise track for platforms and catalog owners requiring tailored integration. It also co-develops custom watermarking and verification solutions with partner organizations and integrates its technology into existing production, distribution and moderation workflows on a longer-term collaborative basis.
Traction
Reported headcount of 1-10 employees. The company states its deepfake detection model has won multiple awards in international audio deepfake detection competitions. It maintains a public GitHub organization with an official Python SDK for its audio watermark API, and was selected for the spring 2025 Techstars Columbus cohort.
Latest developments
Selected in March 2025 for the spring 2025 Techstars Columbus cohort powered by The Ohio State University, described there as creating an AI platform for protecting voices from deepfake manipulation. The company's public Python SDK repository shows continued development activity into 2026.
▸Full profile — market position, technology, go-to-market, geography, history
Market position
Positions itself against academic open-source watermarking research and closed enterprise appliances by offering API-first infrastructure, and publishes commentary on security weaknesses in open-source audio watermarks. It cites awards in international audio deepfake detection competitions and alignment with C2PA and EU AI Act Article 50 as credibility markers.
Stated points of difference are API-first, self-serve delivery with sub-100ms verification latency; identity-based leak matching described as roughly 100x faster than fingerprint comparison, enabling automated web-wide scanning; C2PA-compatible provenance metadata carried inside the audio signal rather than in a sidecar file; watermark imperceptibility validated at PESQ 4.6+ and robustness to compression, editing and re-recording; and frame-level, mobile-capable deepfake detection.
Technology
The watermarking system uses psychoacoustic embedding designed to be imperceptible, with quality validated on ViSQOL and PESQ scores of 4.6 or above at 48 kHz, and is stated to survive compression, editing and over-the-air re-recording. Watermarks carry C2PA-compatible metadata covering ownership, creation timestamp, licensing terms and modification history, embedded in the audio signal itself. Detection uses identity-based matching rather than fingerprint comparison, which the company says is about 100x faster, and verification latency is stated at under 100 milliseconds. The deepfake detection model performs real-time and frame-level classification of human versus AI-generated speech and is optimized to be lightweight enough for mobile devices. Access is via a RESTful API and an official Python SDK published on GitHub.
Go-to-market
Two-track motion: a self-serve developer path where users create an account and obtain an API key without a sales call, and a direct enterprise sales path ("Contact Sales", "Request a Demo") for platform and catalog integrations. The company also pursues partnership-led co-development with rights holders and platforms, and distributes an official Python SDK through GitHub. It participated in the Techstars Columbus accelerator program powered by The Ohio State University in the spring 2025 cohort.
Record labels, music publishers and distributors; streaming services and digital platforms; marketplaces and communication platforms handling large volumes of user-generated content; organizations with strong data security requirements; and AI labs and product teams building generative or voice-driven experiences. A self-serve developer tier targets individual engineers integrating watermarking directly.
Geography
Headquartered in Los Angeles, California (Outpost Estates), and participated in an accelerator program based in Columbus, Ohio.
History
Founded in 2024 and based in Los Angeles. In March 2025 it was named to the spring 2025 cohort of Techstars Columbus, powered by The Ohio State University, a 10-company class spanning health care, fintech, climate and AI applications.
Compiled by commissioned research from 8 cited public sources — announcements, filings, and press listed under research sources below.
Key figures
latest reportedCompany-reported or press-reported figures, each dated to when it was claimed — not independently audited.
Related companies · 1
Companies working in the same space as OfSpectrum.
Timeline · 3
launches, deals, and filingsOfSpectrum maintains a public GitHub organization hosting an official Python SDK for its Audio Watermark API, with repository activity updated in September 2026.
The company states its deepfake detection model has won multiple awards in international audio deepfake detection competitions.
OfSpectrum (Los Angeles) was named one of 10 startups in the spring 2025 cohort of Techstars Columbus, powered by The Ohio State University, described as creating an AI platform for protecting voices from deepfake manipulation.
Dated company events from announcements, filings, and press; legal rows summarize public dockets and regulator releases.
▸Research sources · 8
primary sources listed
- OfSpectrumofspectrum.com · web
8 public sources were cited for this profile; the first-party ones are listed here.
Frequently asked questions
- What does OfSpectrum do?
- OfSpectrum provides inaudible audio watermarking, deepfake detection and provenance APIs for copyright and AI content authenticity.
- Who founded OfSpectrum?
- OfSpectrum was founded by Yechen Wang, Yishu Huang in 2024.
- Who are OfSpectrum's investors?
- OfSpectrum's investors include Techstars.
- Where is OfSpectrum headquartered?
- OfSpectrum is headquartered in Los Angeles, US.