Pre-launch creative scoring
Upload a creative set. Get a 10-point score across six weighted dimensions before any spend. Sample size and benchmarks shown inline on every read.
OmnisciaScore your ad creatives before you spend.
Score creative quality across Meta, Google Ads, and TikTok. Catch clustering risk before delivery throttles. Read patterns no single dashboard can see.
Live from the platform
Every number on this page is queried from Postgres on a 5-minute cache. Updates as the platform grows.
···
Creatives analysed
···
Hours of video transcribed
···
Audio features extracted
···
Briefs generated by Forge
···
Cortex training records
···
Signal articles indexed
···
Score categories
···
Publishing platforms
What Omniscia does
Three core reads, all running before your budget is exposed.
Clustering detection
Spots creatives sharing visual style, audio fingerprint, or hook structure before Meta, Google, and TikTok delivery groups them and throttles. Hook fingerprints plus CLIP embeddings make the overlap explicit.
Cross-platform correlation
One Cortex model reads ROAS, CTR, and CPA across all three platforms. The same creative running on each one means triple the data to learn from.
Nine products. One engine. One closed loop.
Signal and Intel prime the loop. Forge through Scia iterate creative production. Cortex sits over everything, reading from every product and sharpening every score. Hover any node to see how its data flows.
- CortexReads from every product (Signal trends, Intel patterns, Lens scores, Launch outcomes, Tempo deltas, Nexus correlations, Scia syntheses) and conditions every product. Retrains on your linked accounts; every score sharpens.
- Signal27 industry sources, refreshed continuously. Trends Signal surfaces are searchable from Forge briefs and feed Cortex as long-tail context. Signal primes the loop; it does not sit inside it.
- IntelMeta Ad Library plus deep analysis on every competitor you watch. Findings flow into Forge briefs as priors, surface in Scia recommendations, and feed Cortex as cross-account performance signal.
- ForgePulls Signal trends, Intel patterns, Nexus performance reads, and Cortex priors into every directive. Each accepted directive stamps the creative envelope so the next analysis knows the brief that birthed it.
- LensEvery uploaded creative scored across six weighted categories. Score, fingerprint, and CIE envelope stamp onto the asset. Cortex retrains on every linked outcome the score predicted.
- LaunchCross-platform publishing across Meta, Google Ads, and TikTok. Deploys with Cortex-informed priors (audiences that have worked) and stamps the lineage so post-launch performance ties back to its brief.
- TempoScore trends, velocity, and percentile benchmarks over time. Surfaces fatigue 5–7 days early on linked accounts. Feeds Cortex as time-series performance signal so future Lens scores carry sharper benchmarks.
- NexusROAS, CTR, and CPA correlated across Meta, Google Ads, and TikTok. The same creative on each platform is triple the data. Reads aggregate back into Cortex so the engine knows what good looks like by vertical and platform.
- SciaReads Tempo trends, Nexus correlations, Forge directives, and Cortex priors to synthesise the next playbook. Scia closes the loop — its output is what Forge briefs cite next round.
How we're different
We analyse before you spend
Every other platform tells you what went wrong after the budget is burned. Omniscia reads the creatives up front and flags clustering and fatigue risk before any spend.
We read across all three platforms
Meta, Google Ads, and TikTok unified in one Cortex model. Fingerprints match the same ad across platforms. One read instead of three siloed dashboards.
We don't promise your outcome
Scores carry sample size and confidence inline. Fatigue flags are pattern reads, not guarantees. Every cited number names the account count it's drawn from.
Every analysis sharpens the engine
Cortex retrains weekly on anonymised outputs from every linked campaign. After 30+ pairs, your weights blend 70% your data with 30% global.
Customer stories
First cohort, in progress
Omniscia is rolling out to its first cohort of customers right now. As they run real campaigns through Lens, Tempo, Forge, Launch, and Cortex, the stories that appear here will be theirs — with their names, their numbers, and their words, published only with their explicit consent.
We will not anonymise random user data and frame it as a case study. If a number appears on this page, a real customer agreed to it.
How we compare
Most ad creative tools fit one of three buckets. Omniscia spans all three with a single Cortex engine and adds primitives no category currently has.
Variant generators
AdCreative.ai · Pencil · Hawky
Strong on producing variants. Light on cross-platform performance reads and per-account retraining.
Creative dashboards
Motion · Foreplay · Atria
Strong on post-launch reporting and inspiration. Light on pre-launch scoring and Google Ads coverage.
Performance optimisers
Madgicx · Smartly
Strong on bid automation and enterprise delivery. Light on creative diagnostics and competitor scoring.
Your data, protected
TLS in transit, AES-256-GCM at rest, zero-retention API calls to every AI provider. Cortex trains only on anonymised outputs, never your creatives. GDPR-compliant by design.