What is your accuracy rate?

Not published yet, deliberately. Every assessment is logged, timestamped and tracked forward against the market, but the Technical engine is mid-recalibration — a hit-rate measured against a build being replaced would be marketing, not evidence. The number goes on the site with its sample size and forward window once calibration closes.

So why should I trust it now?

Not on faith. Every score decomposes into visible inputs, the rules are written down, and the same inputs reproduce the same number tomorrow. Judge the reasoning during the trial on tickers you already understand — that is a stronger test than a percentage on a landing page.

Is this a signal service?

No. Omnisc produces a direction, a named setup and a written argument. It never tells you to buy or sell, never sets a price target, and is not a registered adviser or broker-dealer.

Can the AI hallucinate a score?

The AI only extracts structure from language — catalyst type, scope, evidence strength. Deterministic formulas turn that structure into every number you see. A model cannot invent a score, only mis-classify an article, which the ranking and decay rules bound.

I have TradingView. Why this?

Keep it. TradingView draws the chart brilliantly and then leaves the analysis to you. Omnisc does the analysis across news, price structure and crowd attention, and hands you the reconciled read. They sit next to each other.

Yahoo Finance is free.

Yahoo lists headlines. It does not tell you which of today's articles moved the assessment for this ticker, or how that interacts with a deteriorating trend. Listing is not triage.

Bloomberg does all of this.

It does, better, for roughly two hundred times the price and with an onboarding process built for institutions. Omnisc is the retail-scale version of one specific Bloomberg habit: reconciling conflicting evidence into a stated view.

Can't I just ask ChatGPT?

You can, and it will answer confidently with stale training data and a different answer tomorrow. Omnisc reads today's articles, applies one fixed rule set, and produces a score you can re-derive next week.

How many articles do you read?

Every article that crosses our news feeds for a ticker — and for each one, the article body itself, up to the first 6,000 characters, not just the headline. They are ranked by relevance, source weight and impact, and the top 50 feed the aggregate score. The rest stay logged.

Can good news flip a bearish chart?

No, and that is a stated rule rather than a preference. Technicals set the base direction. News is a bounded modifier that can shade a call but never flip it, and it is deliberately asymmetric because negative news moves price harder.

Why does Crowd Focus have zero weight?

Because retail sentiment is information, not a signal. It answers whether a move is already crowded, and it sits beside the call rather than inside it.

How often does everything update?

News every 10 minutes during market hours and every 15 minutes off-hours. Technicals daily after the close. Crowd every 3 hours. Fusion several times per trading day. Every card shows a freshness badge.

What happens on weekends and market holidays?

Technical Analysis and Fusion do not recalculate when the market is closed — there is no new price data. News and Crowd keep running. Every card carries a freshness badge, so you always know how current each layer is.

Which tickers are covered?

150+ curated names across AI & Tech, Defense and Materials — chosen because news measurably moves them. Sector spillover means articles about uncovered companies still reach the tickers they affect. Coverage expands toward 200+ in 2027. Subscriber requests are weighted first in the queue.

Can you add my ticker?

Ask. Subscriber requests are weighted first in the expansion queue, and the constraint is data quality per name rather than a hard limit.

Will I be charged during the trial?

No. The card is saved to start the trial and the first charge lands on day 8. Cancel before then and nothing is taken. Annual plans are refundable for 14 days after the first charge.

Do you sell my data?

No. There is no ad model and no data resale. Card details never touch our servers — the Merchant of Record handles payments end to end.

Is this really one person?

The product is. One engineer builds and runs every engine, the Fusion layer and the infrastructure. Marketing and communications is Viktoriia Zybalo. Nobody else touches the scoring rules. The trade-off is stated rather than hidden: slower feature velocity, but support answered by the person who wrote the code and a roadmap that isn't a committee document.

What happens if you stop?

Subscriptions are month to month and cancel in one click, so the exposure is one month of lunch money. On annual billing, the 14-day refund window applies after the first charge. The company is a registered Wyoming LLC with payments handled by an external Merchant of Record, not a hobby account.

Is the product finished?

No. It is in Early Access, live since April 2026, with the Technical engine being recalibrated, an options flow layer in build, and filings, ML calibration and Copilot on the published roadmap. The Stage page lists what works today and what does not.

Can I influence what gets built?

Yes, more than at a larger company. Subscriber requests are weighted first in the coverage queue, and feature suggestions go straight to the person writing the code rather than into a product backlog.