Multi-modal detection, plus calibrated eye tracking and a price tag.
Lightscreen's voice AI interviewer Ethan runs credible multi-modal pattern detection (keystrokes, audio, video, live conversational probing). What's missing publicly: calibrated per-session eye tracking, browser lockdown, any compliance attestation, and (as of 2026) any pricing on the website. Evy ships layered anti-cheat plus $2–4/interview published on /pricing, with real paying customers behind it.
Calibrated eye tracking on top of pattern detection
Lightscreen's multi-modal detection (keystrokes, audio, video pattern analysis, live conversational probing) is a credible way to catch obvious tells, and their CTO has confirmed they added statistical and conversational cheating-detection methods in 2025. Evy layers per-session calibrated eye tracking on top: a gaze model tuned to each candidate at the start of the interview, specifically designed to catch the subtler off-screen reading that pattern detection alone tends to miss.
Transparent $2–4/interview vs enterprise-only
Lightscreen launched in Nov 2024 with public per-minute rates ($0.25/min behavioral, $0.75/min technical), but removed self-serve pricing in 2026. The site now says 'Enterprise Pricing: Talk to us.' Evy publishes pay-as-you-go ($4), monthly ($3), and annual ($2) on /pricing today, so you can pilot in an afternoon without booking a meeting.
Production maturity and compliance vs YC-stage
Lightscreen is YC F24 (Nov 2024), ~$500K raised, ~4 employees, self-reported ~$440K ARR per Getlatka, with no SOC 2, no GDPR posture, no ISO 27001, no HIPAA (even while targeting healthcare hiring), and zero third-party reviews on G2, Capterra, TrustRadius, or Trustpilot. Evy is live with paying customers, real interview volume, and a longer track record of catching the cheating teams actually see in 2026.
FEATURE BY FEATURE
Feature-by-feature comparison
How Evy and Lightscreen stack up on the things buyers ask about.
Teams that want layered anti-cheat plus production maturity
Lightscreen
Early adopters comfortable buying from a four-person YC-stage team
ANTI-CHEAT BREAKDOWN
An interview you can cheat on isn't a screen.
The #1 thing Evy gets right and most AI interviewers get wrong: catching the candidates who run ChatGPT in another window. Here's the honest comparison.
Lightscreen's approach
Multi-modal pattern detection (keystrokes, audio, video, live conversational probing) with undisclosed methodology
Lightscreen's anti-cheat is the most credible part of the product. Their public stack includes behavioral pattern analysis, full keystroke logging, video and audio capture, plagiarism detection, and live conversational probing (the AI interviewer 'Ethan' asks follow-up questions that force candidates to explain their reasoning, which is genuinely hard to fake by pasting from ChatGPT). In 2025 their CTO publicly added 'statistical and conversational methods for AI-cheating detection,' though he explicitly declined to share how they work. That's the gap: the methodology is opaque. There's no published whitepaper, no peer-reviewed study, no audited cheating-rate dataset, no detection-accuracy or false-positive rate, and no third-party review presence (no G2, Capterra, TrustRadius, or Trustpilot). Absent from public materials: no per-session calibrated eye tracking, no browser lockdown, no formal identity verification, and no compliance attestations (SOC 2, GDPR, ISO 27001, HIPAA) even though Lightscreen explicitly targets healthcare hiring. Lightscreen's multi-modal pattern detection is credible. What's missing publicly is calibrated eye tracking, browser lockdown, identity verification, and any published methodology or audit.
Evy's anti-cheat stack
Per-session calibrated eye tracking
Calibrated for each individual candidate at the start of the interview so the gaze signal is meaningful, not a generic vision model. Detects when candidates look off-screen at a second device, a phone propped on the desk, or notes on another monitor. This is the layer Lightscreen's published methodology does not advertise.
Screen + camera recording
Full screen capture of what the candidate sees and full camera capture of what they do. Every interview, on record, exportable for review. You don't get a probability score, you get the evidence and can replay the exact moment a flag fired.
Full-page activity detection
Catches tab switches, copy-pastes, and window blurs (the actual signals of someone running ChatGPT in another window). Each event is timestamped and tied to the question being answered, so reviewers can pinpoint where signal quality dropped.
Lockdown browser mode
Restricts other apps and tabs during the interview so the candidate stays in focus. Optional per role: turn it on for senior or technical screens where the stakes justify it, leave it off for high-volume frontline roles where friction matters.
Layered with pattern detection, not instead of
Evy doesn't replace Lightscreen's approach, it extends it. The same keystroke, audio, and video signals are part of Evy's pipeline, and calibrated eye tracking sits on top as an additional layer. The point isn't that pattern detection is wrong; it's that pattern detection alone misses the candidate who reads quietly off a second screen.
“Anti-cheat that catches the obvious tells is good. Anti-cheat that catches the subtle ones, and tells you how it works, is better.”
PRICING SHOWDOWN
Pricing, at the volumes that matter
Lightscreen launched in Nov 2024 with public per-minute pricing ($0.25/min for behavioral 'Coffee Chats', $0.75/min for technical assessments, plus 1,000 free minutes on signup), but removed self-serve pricing in 2026 in favor of enterprise sales. The site now reads 'Enterprise Pricing: Flexible plans designed to scale with your hiring needs. Talk to us.' With ~$500K raised, ~4 employees, self-reported ~$440K ARR per Getlatka, and the removal of the self-serve tier, the commercial motion is now demo-led. Expect a sales cycle and a custom quote before you see a number.
Evy's transparent pricing lets you pilot in an afternoon: no demo form, no team-size question, no two-week procurement loop. Lightscreen requires a sales conversation before you can see what an interview costs, and the removal of their self-serve tier in 2026 suggests the price has moved up, not down. If you'd rather just try the product, that's the difference.
What teams say when they pick Evy over Lightscreen
“
Evy caught a candidate cheating that we 100% would have moved forward. That alone paid for the year.
VP of People“
The recordings and transcripts mean our hiring manager actually reviews everyone instead of skimming resumes.
Talent Partner“
Reduced our cost-per-hire by ~40%. Hard to argue with the math.
VP of Talent
QUESTIONS
Common questions about Lightscreen vs Evy
Is Lightscreen better than Evy for technical screening?
Lightscreen's original Nov 2024 positioning was 'uncheatable tech screens' for software engineers, and they still ship credible technical screening (multiple coding languages, React/front-end debugging, real-time collaborative coding with the AI interviewer 'Ethan'). Evy works for any role: technical, non-technical, frontline, or executive, with conversational video and adaptive follow-ups that don't require a built-in IDE. Both platforms take anti-cheat seriously: Lightscreen with multi-modal pattern detection and live conversational probing, Evy with that plus per-session calibrated eye tracking and lockdown browser. If your hiring is 100% software engineering with live coding and you're comfortable on a four-person YC-stage team, Lightscreen is a defensible pick. If your roles are mixed, or you want a layer of detection that catches off-screen reading, or you need a compliance posture, Evy is the safer call.
Does Lightscreen have eye tracking?
Lightscreen's public methodology centers on multi-modal pattern detection: keystroke logging, audio analysis, video feed pattern detection, plagiarism detection, and live conversational probing (the AI interviewer asks follow-ups that force candidates to explain their reasoning). They added 'statistical and conversational methods for AI-cheating detection' in 2025, though the CTO has publicly declined to share how those work. What's not advertised on lightscreen.ai: per-session calibrated eye tracking. Calibrated eye tracking is different from generic gaze detection because it builds a per-candidate baseline at the start of the interview, then flags deviations specific to that individual, which catches subtler off-screen reading that a pattern model trained on aggregate data tends to miss. Evy publishes calibrated eye tracking as a named component of its anti-cheat stack and uses it on every interview.
Is Lightscreen production-ready?
Honest answer: Lightscreen is Y Combinator Fall 2024, raised ~$500K in December 2024 from Y Combinator, Rebel Fund, Exitfund, iSeed Ventures, and Llama Ventures, and per Getlatka has ~4 employees and ~$440K self-reported ARR as of Sep 2025. They have strong founders (CEO Prachie Banthia was VP of Product at AssemblyAI, head of product at rideOS which Gopuff acquired for $115M, and a Google APM; CTO Gavin Saldanha was at Cauzal AI and Google Cloud and holds degrees from UC Berkeley and Columbia). The product won #10 Product of the Day on Product Hunt in November 2024 and was named to the GSV Cup 50 in January 2025. But the homepage's '20+ companies' claim isn't independently verified, the four visible logos (Artisan, Sybill, Tennr, Wanderlog) are all YC-backed AI peers, there are no published case studies, no compliance attestations, and zero third-party reviews on G2, Capterra, TrustRadius, or Trustpilot. Evy is further along on production maturity, customer reference, and compliance posture, which matters when you're trusting a vendor with hiring decisions.
How does Lightscreen's anti-cheat compare to Evy's?
Lightscreen runs multi-modal pattern detection: behavioral pattern analysis, full keystroke logging, video and audio capture, plagiarism detection, and live conversational probing where the AI interviewer 'Ethan' asks follow-up questions that force candidates to explain their reasoning. That last one is genuinely effective against paste-from-ChatGPT attacks. In 2025 their CTO added 'statistical and conversational methods' for cheating detection (he declined to elaborate). What's absent from their public materials: per-session calibrated eye tracking, browser lockdown, formal identity verification, any published whitepaper or audit, and any compliance attestation. Evy runs the same multi-modal signals and adds calibrated eye tracking, lockdown browser, and named full-page activity detection (tab switches, copy-pastes, window blurs) on top. Both companies are serious about anti-cheat. Evy's stack is layered (pattern detection plus calibrated gaze plus lockdown plus published methodology) rather than relying on pattern detection alone.
What is Lightscreen's pricing?
As of 2026, Lightscreen does not publish self-serve pricing. The site says 'Enterprise Pricing: Flexible plans designed to scale with your hiring needs. Talk to us.' They previously (Nov 2024) published per-minute rates: $0.25/min for behavioral 'Coffee Chats' and $0.75/min for technical assessments, plus 1,000 free minutes on signup. Those rates have been removed. The shift to enterprise-only suggests the company moved up-market and that the self-serve PLG motion didn't convert at scale. Evy publishes every rate on /pricing: $4 per interview pay-as-you-go, $3 per interview on a monthly seat, $2 per interview on an annual seat, with no commitment required to start.
Did Lightscreen pivot from technical screening to a broader product?
Yes, and they're explicit about it. Lightscreen's Nov 2024 positioning was 'Uncheatable tech screens driven by an AI interviewer.' By mid-2025 the tagline had changed to 'The AI recruiting pipeline that does it all,' and the product was positioned across four verticals: Frontline (warehouse/retail), Healthcare, Tech, and Call Center / BPO. They launched behavioral 'Coffee Chats' interviews in April 2025 and a React/front-end engineering interview type later that year. The pivot makes commercial sense for a YC-stage company chasing a larger TAM, and it's also a signal the company is still figuring out exactly where it wants to compete (notably with ~4 employees stretched across four verticals, including healthcare without any HIPAA attestation).
Can I use Lightscreen for non-technical hiring?
Now, yes. Lightscreen's 2025 pivot broadened the product into frontline, healthcare, tech, and call center solutions, so non-technical roles are explicitly on the menu. Worth noting: the product was originally engineered around technical screening (live collaborative coding, code-context evaluation), so the non-technical surface area is newer. Evy was built from the start to work for any role: sales, ops, customer support, frontline, executive, or technical, with conversational video and adaptive follow-ups that don't assume a coding interview format. If you're hiring across mixed roles in 2026, Evy has more mileage on non-technical interviews.
See the difference in 2 minutes.
Talk to Evy yourself. No demo form, no team-size question, no sales cycle. Just an actual interview so you can see what layered anti-cheat feels like.
