Helix Raises $6.2M Pre-Seed for Voice Verification That Never Stores Your Voice

Helix has raised $6.2 million in pre-seed funding to build a voice-verification platform that confirms a human speaker, estimates age, and returns biometric authentication from four spoken numbers in under four seconds — all on the device, with no audio ever uploaded.

The company announced the round and the launch of its platform with early partners on November 11, 2026. Chicago Ventures led the financing. BAM Ventures, Aglaé Ventures, Corazon Capital, Tusk Ventures, Yonder Ventures, Rule30, Press On Ventures and Brown Angel Group also participated.

The bet is straightforward. Verification today often means collecting exactly the material platforms should least want to hold. CEO and co-founder Alex Oberg framed the problem directly: "The traditional approach has been to collect reams of sensitive data, ranging from uploaded driver's licenses to video selfies." That model, he argues, burdens users while leaving platforms uncertain whether an account belongs to a real adult or a synthetic agent — a question that has grown sharper as AI-generated voices and agents proliferate.

Who Is Behind Helix?

The founding team comes from companies that operate large consumer platforms with acute trust-and-safety exposure. Helix was founded by engineers from Ticketmaster, Tinder and Zillow.

Two industry names anchor the company's governance and advice. Jared Smith, former Ticketmaster president and global chairman, sits on the board and invested. Dag Kittlaus, co-founder of Siri and Viv, serves as an advisor.

How Does the Voice Check Work?

The documented age-assurance workflow keeps every computation local. The system extracts formants (the resonant frequency bands of speech), pitch and vocal-tract resonance directly from the user's speech on the device. Raw audio is immediately discarded. An on-device model estimates age without a server round trip.

The session then returns a signed pass/fail proof. No biometric vector is retained. No recording is uploaded. Across both biometric multifactor authentication and age assurance, Helix states that no biometric data or other personally identifiable information is stored on any server.

That architecture is the product's central distinction from document- and selfie-based checks. A platform can make an age or authentication decision without taking custody of an identity document, a selfie or a voice recording. In an environment where stored identity material becomes breach liability and regulatory exposure, custody-free verification is the pitch.

Helix's product site also positions the same check for repeated use — at a new device, a payout or a password change — extending verification beyond the initial sign-up moment and into ongoing account security.

What Are the Reported Numbers?

Helix reports two headline performance figures:

• 96.8% accuracy at an 18+ age gate for its age-assurance product.
• 99.6% detection of AI-generated voices produced by the newest voice engines.

The company's product page does not supply the underlying test methodology for either figure, which limits what outside evaluators can currently verify. The numbers are vendor-reported until independent testing or published benchmarks appear.

Integration is available through a software development kit (SDK). A single-line integration option exists for Claude Code and OpenAI Codex deployments, and one SDK covers three functions: bot management, biometric multifactor authentication and age assurance.

What Certifications Does the Platform Hold?

Helix claims compliance with three frameworks:

• ISO/IEC 27566-1:2025 — the published ISO standard establishing a framework for age-assurance systems and describing core characteristics, including privacy and security, for age-related eligibility decisions.
• Full accreditation by the Age Check Certification Scheme.
• Certification under the IEEE Online Age Verification Certification Program, which assesses the design, specification, evaluation and deployment of age-verification systems against IEEE 2089.1.

The regulatory context matters here. IEEE connects that certification framework to New York's SAFE for Kids Act, which it says requires covered platforms to independently certify their age-assurance systems. Age-verification mandates of this kind are spreading across jurisdictions and sectors, and platforms facing them need compliant mechanisms that do not multiply their data liabilities. Helix's certification stack is aimed squarely at that procurement requirement.

Who Is Deploying It?

Zeel, an in-home health and wellness platform, is the named early partner. Zeel is testing Helix in beta as an additional verification measure within its trust-and-safety process. CEO Samer Hamadeh described the engagement in deliberately measured terms: "Helix is one of the technologies we're exploring as we think about how Trust and Safety needs to evolve for the AI era."

The phrasing signals evaluation rather than adoption at scale. The beta's value to Helix is as an early test of how a signed, on-device result fits an operating trust-and-safety workflow in a real business.

Where Could This Land Next?

Helix identifies five potential deployment sectors: ticketing, dating, gaming, gambling and fintech. All five share a structural problem — verifying that a user is a real adult human, repeatedly and across devices, without accumulating sensitive identity stores that regulators and attackers both target.

The open business question is whether a signed on-device result can satisfy the different risk and compliance needs of those markets without creating another repository of sensitive identity material. The certifications and reported performance figures support Helix's case for that result. The Zeel beta supplies the first evidence of how the approach behaves inside a live trust-and-safety operation.

The pre-seed stage sets expectations. A $6.2 million round funds a product and a handful of pilots, not a market. What it buys is time to prove that four spoken numbers, processed on a phone and reduced to a cryptographic pass/fail, can stand in for the driver's licenses and video selfies that Oberg says the industry has relied on for too long.