Common questions

No. Trades are stored on your device in a local SwiftData database. When iCloud sync is on, SwiftData replicates the database through CloudKit's user-private storage — which lives in your iCloud account, not in any Kyra-operated service. Kyra has no backend, no analytics pipeline that sees trade content, and no third-party data processors with access to your history.

For the full architecture, see Why on-device.

When iCloud sync is on, your trades, settings, and checklist sessions sync through CloudKit. The data lives in your personal iCloud — the same storage that holds your Photos, your Notes, and your other Apple-mediated app data. Kyra cannot read your iCloud; Apple's CloudKit doesn't grant developer-side access to user-private data.

If you turn iCloud sync off, trades remain on the single device you logged them on. You can switch sync on and off at any time in Settings.

Yes. Kyra is built to work fully offline. Trade logging, history browsing, pattern detection, the pre-trade checklist — all of it runs locally. The only time the app needs network access is when iCloud sync runs in the background (handled by Apple, not by Kyra) and when you fetch a software update through the App Store.

Settings → Privacy & Security → Privacy Lock. Turn it on, and Kyra requires Face ID, Touch ID, or your device passcode before showing any trade data. The lock kicks in when you open the app and when you bring it back from the background.

Privacy Lock also blurs the app in the iOS app switcher so trade information isn't visible during a swipe-up gesture in public.

If iCloud sync is off, deleting the app deletes the local database — the trades are gone unless you exported them first.

If iCloud sync is on, the data persists in iCloud. Reinstalling Kyra and signing into the same Apple ID restores the full history.

Kyra collects three things: Device ID (anonymous, used by TelemetryDeck for usage analytics), Product Interaction (anonymous, what features you tap), and Crash Data (anonymous, used for stability fixes). None of it is linked to your identity. None of it includes trade content. There is no cross-app tracking SDK.

You can turn analytics off in Settings → Privacy & Security if you want zero data sent.

The first Tracking-tier candidates appear within the first 5–10 trades. Hint-tier patterns typically appear around 30–50 logged trades. Signal-tier patterns typically appear around 75–150 logged trades. Proven-tier patterns require at least 30 trades that match a specific condition, which usually means 200–400 trades in total.

The sample sizes vary because patterns are personal. A trader whose emotional baseline is highly varied will get patterns sooner than a trader whose trades are nearly uniform. The engine waits for enough evidence before claiming.

Four tiers, in order:

• Tracking — too few trades to claim a pattern; the engine surfaces a ghost card noting the candidate
• Hint — pattern is plausible; uncertainty range is still wide
• Signal — pattern is meaningful; uncertainty range has narrowed enough to be informative
• Proven — pattern is robust; uncertainty range is narrow enough to anchor a decision

Each promotion requires the pattern to pass a significance gate that controls for multiple-comparison error. See Pattern detection for the methods.

No. Kyra surfaces statistical observations from your past trade history. Past performance does not guarantee future results, and the engine does not extrapolate. The patterns it surfaces tell you what your trades have done historically under specific conditions, not what they will do.

Kyra uses on-device mathematical AI — classical statistical inference run on your trade history, in milliseconds, on your phone. It is not LLM-based and there is no chatbot. The methods belong to a long-established class of inference that fits small data exceptionally well.

Why this class of inference fits the problem:

• Sample sizes in a personal trade history are small (typical user has 100–500 trades); classical inference outperforms LLMs and large ML models on small samples
• The math is auditable end-to-end — every claim has an uncertainty range the user can read
• Sample-size honesty is structural — the uncertainty range widens at low sample size by construction
• The math runs on-device; nothing needs to be uploaded for the engine to work

Kyra does not upload your trade history to a language model or present itself as a chatbot.

The methods are named and the math is explained on The math behind Kyra.

"Accurate" applies to predictions; the engine does not make predictions. The right question is "how trustworthy are the surfaced claims" — and the answer is: every claim has an explicit uncertainty range. A Signal-tier pattern surfaced as "Anxious trades have lost 23% more than baseline (range: 11% to 36%)" means exactly that — given your history, the gap most likely sits inside that range.

If the engine could be confidently more precise, the range would be narrower. It isn't artificially tight to look impressive.

Quick P&L is the three-field fast-logging mode: direction, P&L (a single dollar amount), and the emotion you felt. No ticker required, no price math. Under 10 seconds per trade.

Trades logged via Quick P&L feed the same pattern engine as full-entry trades. The engine cares about emotion and execution rating more than it cares about exact prices.

Yes. Tap the trade in History to open the detail view. Tap Edit. Change any field. Save. Pattern detection recomputes on the next view.

No. Trades are entered in Kyra; it does not import broker files or connect through OAuth.

Avoiding live auto-sync means trade data does not pass through a Kyra-operated server. See Why on-device.

Yes, for single-leg options. Kyra's setup tags include "Earnings Play" and "Scalp." Multi-leg structures (verticals, condors, butterflies) are recorded as single P&L entries — Kyra captures the outcome and emotional state, not individual legs.

Streaks are gamification. Gamification rewards engagement; trading rewards judgment. The two often conflict — a "don't break the streak" pressure can push a trader to take an avoidable Friday-afternoon trade. The damage to the account exceeds whatever the streak was worth.

Kyra deliberately does not ship streaks, badges, level-ups, or any other gamified mechanic. See Manifesto.

Free includes unlimited trade logging, full filterable history, and your single strongest detected pattern. Free is fully functional as a journal.

Premium ($4.99/mo or $39.99/yr) adds every pattern Kyra finds, detailed pattern breakdowns, and the adaptive pre-trade checklist that personalizes to your detected blind spots.

Full pricing breakdown lives on Pricing.

iOS: Settings → Apple ID (top of Settings) → Subscriptions → Kyra Trading → Cancel Subscription.

You can cancel anytime. The cancellation takes effect at the end of the current billing period — if you cancel on day 20 of a monthly cycle, you keep Premium through day 30, then revert to Free.

Refunds are handled by Apple at reportaproblem.apple.com. Apple typically issues refunds within 24–48 hours when the request fits their policy. Kyra does not have a refund system of its own because it has no payment processor — Apple handles billing end to end.

Family Sharing is not currently enabled for Kyra. Each Premium subscription is tied to the individual Apple ID that purchased it.

Kyra is built for iPhone around local storage, on-device pattern detection, and private iCloud sync without a Kyra account or backend. There is no Android app.

The architecture is the headline difference. Cloud-based journals store trades on company servers and often connect to brokers through OAuth. Kyra stores trades on your device and uses classical statistical methods for pattern detection.

The practical consequences:

• Your trade data never leaves your phone with Kyra; it always does with cloud journals
• Kyra has no account system; cloud journals require email + password
• Kyra surfaces patterns with explicit sample sizes and uncertainty ranges; cloud journals often do not
• Kyra works fully offline; cloud journals depend on a network

The trade-offs: Kyra is iOS-only, has no web app, and does not offer live broker auto-sync. Each is a deliberate constraint, not a missing feature.

Mechanically: yes. Kyra runs classical statistical inference — a form of mathematical AI — on-device, on your own trade history. The engine is auditable and the methods are explained on The math behind Kyra.

Kyra is not an LLM chatbot or trading adviser. It helps you read historical patterns; it does not tell you what to trade next.

No. Kyra does not connect to brokers or provide live auto-sync. You enter completed trades in the app; trade data does not pass through a Kyra server.

Kyra is built for Apple's local-data stack: SwiftData on the device and private CloudKit sync through your iCloud account. That combination supports an iPhone journal without a Kyra account or backend. There is no Android or web app.

A web app would require a server to hold the trade data. That conflicts with the architectural commitment that Kyra has no backend. See Why on-device.

If web access is a hard requirement for your workflow, Kyra is not the right journal. Several cloud-based competitors offer web apps and would be better matches.

No. Some apps in this category (Profit AI, for example) accept a chart upload and return a verdict on whether the setup matches your saved strategy rules — Ideal, Good, Not Ideal, or Poor. The verdict is a recommendation on a specific possible trade.

Kyra does not give per-trade verdicts. It presents observations about your past trades, never a recommendation on the next one. See Manifesto Section 2 (Not a recommendation engine) and the comparison page at Kyra vs Profit AI.

Not as a designed-for use case. Kyra's persona is the retail self-executor — a trader running their own equities or options account, typically with a personal brokerage. Apps built around prop firm workflows (Profit AI, others) ship specialized configurations for evaluation phases, funded-account daily-loss limits, and prop firm sub-account mapping. Kyra does not.

If you are running an FTMO or TopStep workflow as your primary trading activity, the prop-firm-shaped apps will fit your workflow better than Kyra will. If prop firm trading is a side activity and your primary trading is your own account, Kyra works fine — it just does not surface prop-firm-specific UX.

For anything not covered here, the support page has a direct contact path.