Terms of Service, Privacy Policy, Risk Disclosure, AI Disclosure and Marketplace terms. These are working drafts for a preview build and must be reviewed by qualified counsel in each target jurisdiction before launch.
Scope of service. AI Finance Hub provides market data, analytical tooling, strategy infrastructure, marketplace distribution and optional execution technology. We are not a broker, dealer, custodian, investment adviser or portfolio manager. We do not hold client funds and we do not execute orders on our own account.
Account responsibility. You are responsible for the accuracy of your account information, the security of your credentials and every order placed through your account, including orders placed by automated components you have enabled. Enabling automated execution is your decision and remains under your control at all times.
No advice. Nothing on the platform constitutes investment, financial, tax or legal advice, or a recommendation to buy or sell any instrument. Analytical output, including AI-generated analysis, is information for your own evaluation.
Acceptable use. No scraping beyond documented API limits, no reverse engineering of proprietary strategy logic or knowledge content, no attempts to bypass authentication, authorisation, risk controls or the kill switch, and no use of the platform for market manipulation or unlawful activity.
Suspension. We may suspend access to protect the platform, other users or the integrity of trading systems. Where a security or trading incident is suspected, suspension may be immediate and without prior notice.
Data we collect. Account and identity data, device and session data, entitlement and billing records, platform telemetry, and the trading and audit records required to operate the service. We collect only what the service needs.
Separation of telemetry and trading records. Product analytics are kept separate from financial and trading records. Trading credentials and sensitive portfolio data are never used as generic analytics attributes.
Credentials. Broker and exchange credentials are encrypted at rest, scoped per account and environment, and are never exposed to the browser or used for analytics. Payment card data is never stored by us; payment methods are tokenised through a PCI-compliant provider.
Retention. Retention periods are defined per data class — account, AI, market and audit data each have their own policy. Certain trading and audit records must be preserved for legal and regulatory reasons and are therefore not deletable on request; where that applies, we say so explicitly.
Your controls. Consent controls where legally required, data export, and deletion workflows that respect applicable law while preserving legally required records.
Trading risk. Trading financial instruments carries a high level of risk and may not be suitable for all investors. You may lose some or all of your capital. Do not trade with money you cannot afford to lose.
Leverage. Leverage magnifies both gains and losses and can work against you as well as for you. In adverse conditions, losses can exceed your initial deposit depending on the instrument and the terms of your broker.
Performance is not predictive. Historical performance, backtested results and any past signals are not a guarantee, promise or reliable indicator of future results. Reproducibility of a backtest means the result can be verified — it does not mean the result will repeat.
Automated execution risk. Automation removes reaction delay and enforces your stated policy. It does not reduce market risk. A configuration error, a connectivity failure, a data gap or a mis-specified rule can cause losses at speed. Automated execution is off by default for this reason.
Third-party products. Marketplace products are supplied by creators or by us as tooling. Installation and configuration are your responsibility. Verification status describes the review a product received, not its profitability or suitability for you.
Data limitations. FX spot volume is decentralised; broker and tick volume do not represent global market turnover. Where a futures contract is used as a proxy for a spot market, this is labelled. Data may contain errors, gaps or corrections, and quality flags are provided so you can assess suitability.
AI is advisory. AI components analyse data, explain methodology and generate candidate ideas. They are reasoning and advisory tools. They do not make binding decisions, do not size positions, do not place orders and cannot bypass the Risk Engine under any circumstances.
Model output can be wrong. Language models can produce confident, well-formatted, incorrect statements. Every AI output should be treated as a hypothesis that requires verification against the underlying data, the strategy version and your own analysis.
Provenance you will see. AI analysis displays its timestamp, data snapshot and time range, model or provider identifier, strategy or rule version where applicable, and a confidence or uncertainty indication where the model can express one. AI sessions, messages and tool calls are logged for audit.
External content is untrusted. News, documents, web pages and text you supply are treated as untrusted input. Prompt-injection defences are applied, tools are drawn from an allowlist and outputs are validated against per-tool schemas before any downstream use.
No guarantee implied. AI analysis never implies guaranteed returns, and we do not present model confidence as a measure of trading probability.
Licence types. Products are offered under purchase, subscription, rental or lifetime licence models, set per listing by the publisher. Licence scope, seat limits and update rights are declared on the listing before purchase.
Server-side enforcement. Licence enforcement is server-side. A product cannot be unlocked by modifying a frontend flag, and entitlement is derived from verified billing state rather than a client-submitted value.
Verification, not endorsement. Creator submissions pass malware and static scanning, package validation, dependency review, signature and hash generation, sandbox testing, permission declaration and human approval. Verification confirms these steps were completed. It is not a performance claim and not an endorsement.
Immutable versions. Published versions are immutable. Updates create new versions with new hashes and changelog entries. A backtest or test result is tied to the specific version that produced it.
Refunds. Refund terms are published per listing by the publisher and recorded as auditable subscription or order events.
Removal. We may remove a listing that fails re-verification, breaches these terms or presents a security risk. Purchasers are notified and licence handling follows the terms published at purchase.
Where the boundary sits. Providing data, information and analytical tooling is generally treated differently from providing regulated advice, asset management, order execution or solicitation. That boundary varies materially by jurisdiction and by the type of client, and it is not something we will decide informally.
What launch requires. Jurisdiction-specific legal advice, published final terms, confirmed data-provider licences, a reviewed risk disclosure, and a documented determination of whether the product crosses from tooling into a regulated activity in each market served.
What we will not do. We will not market guaranteed returns, imply a regulatory status we do not hold, or present our verification process as a substitute for your own due diligence or for regulated advice.
Enterprise engagements begin with a jurisdiction scoping review. We would rather delay a launch than misrepresent what the product is.