Table of Contents

deepeye by deepidv Investment Report

deepeye by deepidv Investment Report

Category: Deepfake detection, identity verification, fraud prevention, and compliance software

Company Stage: Seed-stage

Founder or Founders: Shawn-Marc Melo, founder and CEO; Omar Tahir, CTO and founding engineer

Headquarters: Conflicting public information: LinkedIn identifies San Francisco; the March 2026 funding announcement describes Toronto as the primary engineering headquarters and San Francisco as the U.S. commercial hub

Funding: $1 million seed round announced in March 2026; investors not disclosed

Business Model: Free consumer browser product supporting usage-based verification APIs, SaaS subscriptions, and enterprise contracts

Product Hunt Launch Date: September 2, 2026

Report Date: September 5, 2026

Investment Metric	Assessment
Venture Potential	69/100
Unicorn Path	Conditional
Valuation Attractiveness	Not Assessable
Evidence Confidence	55/100
Final Decision	Watch

Executive Summary

deepeye is a Chrome extension and WhatsApp-integrated deepfake detector from deepidv. It allows users to examine images, video, audio, and text while browsing rather than uploading files to a separate forensic dashboard. The Chrome extension is positioned for consumers, journalists, researchers, online shoppers, and people evaluating suspicious profiles or messages (Product Hunt, Chrome Web Store).

The more significant investment opportunity is deepidv’s broader B2B platform. The company sells identity verification, document authentication, liveness, deepfake detection, AML screening, background checks, compliance automation, and continuous monitoring through APIs and enterprise workflows (deepidv, pricing). deepeye can therefore operate as a free distribution channel and product demonstration rather than the primary source of revenue.

The strongest positive signal is the combination of a genuine enterprise problem, a funded operating company, frequent product releases, and a potentially high-value usage-based model. The company announced a $1 million seed round in March 2026, although investors and valuation were not disclosed (Biometric Update).

The principal concern is evidence quality. The Chrome extension had only 18 users when researched. The company reports more than two million identities verified and hundreds of companies onboarded, but provides no audited revenue, named enterprise customer set, retention metrics, or independently reproduced accuracy benchmarks. Its own benchmark report claims more than four million synthetic-identity events and 10,000 confirmed deepfakes, despite LinkedIn listing the company as founded in 2025 while the report’s dataset begins in January 2024. The relationship between predecessor data, test data, and live customer traffic is not explained (benchmark report, LinkedIn).

The product is relevant and the broader market can support a venture-scale company. However, deepeye’s consumer adoption is minimal, differentiation is vulnerable to well-funded competitors, and critical commercial and technical claims require verification. The decision is Watch, with an upgrade to formal diligence contingent on independent model testing and verified enterprise traction.

Product Overview

deepeye addresses the difficulty of recognizing AI-generated or manipulated media during ordinary internet use. Instead of requiring users to download content and submit it to another application, the extension lets them select media or text from the current webpage and receive a real-versus-AI result.

The Chrome listing says the product detects AI-generated images, manipulated faces, machine-written text, and misleading photos. Product Hunt additionally describes image, video, audio, profile-photo, voice-note, and video-call detection, with Chrome and WhatsApp delivery. The difference between these descriptions suggests that some capabilities may exist outside the public Chrome extension or remain incompletely documented.

The extension is free, although the Chrome Web Store labels it as offering in-app purchases. Consumer pricing or usage limits are not publicly disclosed. The underlying enterprise API lists deepfake detection at $0.60 per verification (pricing).

The current Chrome product is version 1.0.4, last updated July 30, 2026. It had 18 users and no material review history when examined. It requests or processes broad categories including website content, user activity, communications, authentication information, location, financial information, and personally identifiable information. The developer states that data is not sold and is not used outside approved core-functionality purposes (Chrome Web Store).

The browser-native workflow is convenient, but product accuracy cannot be inferred from usability. No independent test of deepeye’s false-positive rate, false-negative rate, robustness to compression, or performance on newly released generative models was found.

Product-quality assessment: Useful workflow and broad modality coverage, but detection reliability remains insufficiently validated.

Founder and Team Assessment

Shawn-Marc Melo is the founder and CEO. A 2026 interview describes him as having built and exited two previous companies and having worked in sales, financial services, and identity technology. However, the prior companies, buyers, transaction values, and founder proceeds were not independently documented in the sources reviewed (Shift & Thrive interview).

Melo reported that development began in February 2025, that he initially funded the company, and that the team had grown beyond 20 people. LinkedIn showed 19 associated employees and an 11–50 employee size range, broadly supporting—but not independently verifying—the team-size claim (LinkedIn).

CTO Omar Tahir’s public profile shows a computer-science degree from Western University and previous experience as software engineering lead at MarketBox. He joined deepidv in 2025 (Omar Tahir). The funding announcement also identifies Robert Herjavec as a board member, alongside Melo and Tahir (Biometric Update).

The company has credible engineering output across APIs, mobile applications, cryptographic attestations, and an MCP server. Public GitHub repositories were recently updated, but had very low star and fork counts; this supports active development rather than external developer adoption (GitHub).

Founder Assessment: Strong founder intensity and credible technical shipping, but claimed prior exits, enterprise sales capacity, and biometrics research depth require further verification.

Market Opportunity

The initial paying customer is not the consumer installing deepeye. It is a regulated or fraud-exposed organization—such as a bank, fintech, marketplace, gaming operator, hiring platform, or identity provider—that needs identity verification and synthetic-media screening.

The problem is commercially material. Deloitte estimates that generative-AI-enabled fraud could cost U.S. financial institutions and customers as much as $40 billion annually by 2027 under its modeled scenarios (Deloitte). The FDIC reported 4,238 insured U.S. institutions as of August 2026, before counting credit unions, fintechs, insurers, marketplaces, telecommunications companies, and non-U.S. regulated entities (FDIC BankFind).

A reasonable bottom-up scenario is 20,000–50,000 globally addressable organizations with sufficient fraud or compliance volume. At an assumed blended annual contract value of $25,000–$50,000, the addressable revenue opportunity would be approximately $500 million–$2.5 billion. These are analyst assumptions, not company forecasts.

The market can support venture-scale revenue, but established identity-verification companies already have compliance certifications, customer references, integrations, and large training datasets. Market growth therefore does not automatically translate into share for deepidv.

Traction and Growth Signals

deepidv reports more than two million completed identity verifications and coverage across 211 countries. A July 2026 company post claimed “hundreds of companies” had been onboarded since launch. Neither figure is supported by customer-level disclosures or audited reporting (company website, LinkedIn).

Its internal benchmark report says it analyzed more than four million synthetic-identity events and over 10,000 confirmed deepfake attempts. The report publishes claimed multi-layer detection rates above 98% for several attack classes, but does not disclose sufficient independent methodology, holdout datasets, confidence intervals, or third-party replication to treat those rates as validated (benchmark report).

deepeye received 121 Product Hunt points, 17 comments, and ranked 11th for September 2. That shows launch-day attention only. The underlying deepidv Product Hunt profile has 349 followers and one review (Product Hunt launch).

The Chrome extension’s 18 users are a much more conservative signal of actual public adoption. The two iOS applications examined each had only one rating (deepidv iOS console, NFC verification app).

Revenue, ARR, paying-company count, transaction growth, customer concentration, retention, gross margin, and net revenue retention are not publicly disclosed.

Traction Assessment: Product development is active and company-reported usage is substantial, but independently visible adoption and commercial traction remain weak.

Competitive Position

deepeye competes directly with Resemble AI’s free Chrome deepfake detector, which also scans images, audio, and video and provides confidence scores and forensic explanations (Resemble AI). Enterprise competitors include Reality Defender, GetReal Security, Pindrop, and identity platforms such as Entrust/Onfido, Persona, Veriff, Jumio, and Sumsub.

GetReal raised a $17.5 million Series A from specialist cybersecurity investors and combines image, audio, video, and real-time communication analysis, demonstrating that well-financed specialist competition already exists (GetReal). Entrust’s acquisition of Onfido also shows that identity verification is consolidating into broad security platforms (Entrust).

deepidv’s differentiation is breadth: deepeye can feed into identity, KYC, AML, background-check, monitoring, and cryptographic-attestation workflows. Its partnership with Scam.AI adds another detection engine and shared datasets (Biometric Update).

However, the partnership complicates the company’s “no third-party APIs” and “owned end to end” positioning. The exact proportion of proprietary versus partner-supplied data and functionality needs clarification.

If a major identity platform launched the same browser feature within six months, customers would remain only if deepidv delivered demonstrably higher detection accuracy, easier workflow integration, lower cost, or a valuable proprietary fraud dataset. None is yet independently proven.

Defensibility Assessment: Medium-Low

Business Model and Economics

The consumer extension is free and appears intended to create awareness, collect threat signals, or funnel organizations toward the enterprise platform.

deepidv charges a $19 monthly maintenance fee for pay-per-use access after the first month. Public plans include Starter at $299 per month, Growth at $1,499, and Scale at $5,999, implying annual contract values of approximately $3,588, $17,988, and $71,988 before overages. Enterprise pricing is custom, with discounts for annual commitments above $100,000 (pricing).

The usage model is attractive if verification revenue scales faster than data, inference, document-validation, blockchain, support, and manual-review costs. Gross margins and cost per check are unknown. High false-positive rates could also increase customer-review costs and undermine retention.

Enterprise onboarding may be expensive. Melo stated that customers are typically compliance officers and CTOs and that integration can take two to three months (founder interview). That makes efficient implementation and expansion revenue essential.

Unicorn Path

Assuming a 10× ARR multiple for a rapidly growing, high-retention identity-security platform, deepidv would need approximately:

$1 billion ÷ 10 = $100 million ARR.

At the $5,999 monthly Scale price, this equals approximately 1,390 customers. At a $100,000 enterprise ACV, it equals 1,000 customers. A mixed scenario could involve 700 enterprise customers at $100,000 plus approximately 1,670 Growth customers at $18,000, totaling about $100 million ARR.

That outcome requires strong gross margins, low fraud-related liability, independently validated accuracy, enterprise-grade certifications, and repeatable sales into regulated organizations. deepeye alone is unlikely to support a unicorn; it must become a distribution and data advantage for the broader verification platform.

Unicorn Path: Conditional

Valuation Assessment

The company announced a $1 million seed round, but investors, security type, valuation, dilution, and current fundraising status were not disclosed. Revenue and growth are also unavailable.

Valuation Attractiveness: Not Assessable

Assessment requires verified ARR, gross margin, customer retention, customer concentration, burn, runway, cap table, round terms, post-money valuation, and the commercial contribution of deepeye.

Key Risks

Unvalidated accuracy: No independent benchmark establishes real-world false-positive and false-negative performance.
Commercial opacity: Revenue, customer count, retention, and concentration are undisclosed.
Minimal deepeye adoption: The Chrome listing showed only 18 users.
Data inconsistency: The benchmark period predates the publicly listed 2025 founding date.
Competitive intensity: Larger identity vendors and funded deepfake specialists can bundle similar functionality.
Privacy exposure: The browser extension handles sensitive content and multiple categories of personal data.
Regulatory and liability risk: Incorrect verification decisions can create financial, discrimination, and compliance consequences.
Positioning inconsistency: “Owned end to end” is difficult to reconcile with partner modules and integrations.
Long enterprise sales cycles: Two-to-three-month integrations can increase acquisition cost and slow revenue.
Detection-model decay: New generation techniques can quickly reduce model effectiveness.

Final Assessment

Venture Potential: 69/100

Category	Score
Market Size and Expansion Potential	18/20
Traction and Growth Evidence	10/20
Founder and Team	11/15
Product Strength	8/10
Distribution Potential	9/15
Business Model and Economics	8/10
Defensibility	5/10
Total	69/100

The strongest elements are market urgency, a broad enterprise product, usage-based pricing, and active technical execution. The weakest are independently unverified traction, limited deepeye adoption, uncertain detection accuracy, and crowded competition.

Evidence Confidence: 55/100

Funding, public pricing, product availability, team identities, app listings, GitHub activity, and Product Hunt statistics are verifiable. Verification volume, customer count, accuracy, country coverage, certifications, and prior founder exits are primarily company-reported. Revenue, retention, margins, burn, valuation, and detailed financing terms remain unavailable.

Final Decision: Watch

The company is potentially venture-backable, but the available evidence does not yet justify formal diligence. The product’s strategic value depends on whether free consumer distribution converts into proprietary threat data and paid enterprise relationships.

Upgrade Conditions

Upgrade to DD if deepidv demonstrates:

At least $1 million in verified ARR with material year-over-year growth.
Referenceable regulated-enterprise customers.
Independent deepfake benchmarks across unseen, compressed, and adversarial media.
More than 70% six-month customer retention.
Gross margin above 65% after inference, screening, support, and manual review.
Meaningful deepeye adoption, such as 50,000 active users or documented enterprise deployment.
Verified SOC 2 and ISO certification reports.
A clear explanation of data provenance and proprietary versus third-party capabilities.

Downgrade Conditions

Downgrade to Pass if consumer adoption remains negligible, enterprise claims cannot be verified, model accuracy deteriorates outside internal datasets, certifications are unavailable, customer acquisition remains services-heavy, or competitors eliminate the company’s differentiation.

Questions for Further Diligence

What are current ARR, MRR, growth, and revenue by product?
How many paying companies are active, and how many are production enterprise deployments?
What are 90-day and 180-day retention and net revenue retention?
How are the two million verifications and four million synthetic events defined and reconciled?
Why does the benchmark dataset begin in January 2024 if development reportedly began in 2025?
What are deepeye’s active users, scan frequency, free-to-paid conversion, and retention?
What independent evaluations validate false-positive and false-negative rates?
What percentage of detection functionality and data comes from Scam.AI or other providers?
What are gross margin, cost per verification, manual-review rate, and infrastructure cost?
Which SOC 2 and ISO 27001 reports can customers inspect?
What are CAC, sales-cycle length, implementation cost, and payback by segment?
Who invested in the seed round, and what are the cap table, valuation, runway, and current financing terms?