What's New in GazeIQ

Launched the /tools suite: four free utilities for ad creative teams

We shipped a set of free, no-login-required tools to help teams evaluate ad creatives without signing up first. Each tool is standalone, indexable, and links back into the full GazeIQ workflow when you're ready for deeper analysis.

• →Free Ad Analyzer — drag-and-drop attention scoring in under 10 seconds
• →Attention Heatmap Generator — visual overlay for any static creative
• →ROAS Calculator — projects return based on creative attention score + spend
• →Facebook Ad Analyzer — platform-specific scoring calibrated for Meta Feed

Platform-specific benchmark library is live

You can now compare your creative's attention score against benchmark distributions for Facebook Feed and Google Display. We built this from an aggregated sample of scored creatives across the customer base. You'll see percentile placement alongside your score, so a 72 means something concrete ("top 25% for Facebook Feed product ads") instead of an abstract number.

Two new pillar guides shipped

We published two long-form reference guides to help teams build attention thinking into their creative process from brief to post-launch. Both are free, no-login, and linked from every relevant spot in the product.

• →The Ad Creative Testing Playbook — full pre-launch testing workflow
• →Attention in Advertising — the science and application of visual attention for ads

Upgraded saliency model to TranSalNet v2

We rolled out a new saliency backbone that materially improves heatmap accuracy on ad-specific imagery. Correlation coefficient against ground-truth eye-tracking data on SALICON moved from 0.89 to 0.907. You'll notice tighter fixation zones around faces, product hero areas, and CTAs.

• →CC (correlation coefficient): 0.89 → 0.907 on SALICON
• →NSS score improvement: +6.2% on internal ad creative test set
• →No action required — all new scans automatically use the updated model

A/B pre-testing now supports up to 5 variants with radar-chart winner analysis

You asked for a cleaner way to compare variant creatives side by side. We shipped a multi-variant mode that lets you upload up to 5 creatives in a single session and returns a radar chart across six attention sub-dimensions — CTA visibility, headline salience, product focus, visual hierarchy, mobile legibility, and offer clarity — plus an overall ranked winner.

• →Upload 2–5 variants at once
• →Radar chart surfaces which variant wins which dimension
• →Exportable as PNG for creative review decks

Instagram Story and Google Display platform context simulation

Every scan now runs in the context of a specific platform placement, not just a generic preview. We added Instagram Story (9:16) and Google Display (300×250 and 728×90) on top of the existing Facebook Feed simulation. Attention scoring is recalibrated per platform because viewer dwell time and scan patterns differ significantly.

Razorpay billing for India + international plan structure

We added Razorpay as a payment processor alongside Stripe so Indian customers can pay in INR without conversion friction. At the same time, we restructured pricing into three global tiers — Starter at $29/mo, Pro at $79/mo, and Agency at $249/mo — with localized pricing for India.

• →Razorpay for INR billing (UPI, cards, netbanking)
• →Stripe for international cards
• →Starter $29 / Pro $79 / Agency $249 per month

Google Gemini fallback LLM on the free tier

Free-tier recommendations previously ran exclusively on OpenAI. We added Gemini as a fallback path so free-tier users still get AI recommendations when OpenAI rate limits or outages hit. Pro and Agency tiers continue to use GPT-4o as the primary model for higher-specificity recommendations.