Raxx Product Principles — Securities-Law Relevant

NOT LEGAL ADVICE. This document reformats locked product decisions for attorney review. These principles are inputs to the attorney's analysis, not conclusions. The attorney's job is to confirm, challenge, or qualify each principle's legal implication — not to accept it at face value. Prepared: 2026-05-13 UTC.

Purpose

The following principles reflect locked operator decisions about how Raxx works and how it is positioned. They are relevant to the securities-law analysis because each one affects the § 202(a)(11) three-prong test, the IA Act "advising others" element, and the Custody Rule analysis. The attorney is asked to:

(a) Confirm whether each principle is adequately reflected in the product as described (the attorney will receive product access for review upon engagement). (b) Flag any principle whose legal implication differs from how it is stated here. (c) Identify any principle that requires additional UX safeguards, copy changes, or documentation to be legally defensible.

Principle 1 — The user authors every strategy

Statement: Raxx users write their own trading strategies. The platform does not propose, suggest, or rank any strategy. No strategy originates from Raxx.

Rationale: A platform that executes a strategy the user authored is analytically distinct from a platform that advises the user to adopt a strategy. This is the foundation of the "pure parser" argument under SEC Release IA-1092 (1987): the AI performs translation, not reasoning about what the user should do.

Implication for attorney review: Confirm that the MBT UI, copy, and onboarding flow does not create any factual record that Raxx proposes or suggests strategies. Flag any surface where the platform's behavior could be characterized as strategy authorship (e.g., auto-completing strategy parameters, offering default templates without user initiation).

Principle 2 — AI is a parser, not an adviser

Statement: Raxx's AI layer (Claude) translates the user's natural-language strategy description into a structured DSL. Translation is the entire scope of the AI's function at execution. The AI does not score feasibility, rank strategies, recommend approaches, or tell the user whether their strategy is good or appropriate.

Rationale: The statutory "advising" element in § 80b-2(a)(11) requires a communication directed toward a conclusion — a recommendation to act or not act. Pure translation (language A into language B) does not direct the user toward any conclusion about what to trade. The analogy: a word processor formats a legal document but does not practice law.

Implication for attorney review: Confirm the line between translation and gap-filling. If the AI fills in an ambiguous instruction (e.g., user says "30-delta iron condor" but doesn't specify which expiry), and Raxx infers a default, does that inference constitute advice? The attorney should identify where gap-filling, if any, occurs and whether it requires additional user-confirmation UX or explicit user consent.

Principle 3 — Execution is deterministic and rule-based

Statement: Once a strategy is activated, Raxx's execution engine fires orders rule-by-rule, deterministically, against the user's broker account. There is no AI in the order-firing path. The same input always produces the same output. Execution is predictable, auditable, and user-reviewable at every step.

Rationale: The SEC has stated that "an adviser cannot defer its fiduciary responsibility to an algorithm." This concern is most acute when the algorithm makes discretionary decisions. Raxx's execution layer is not discretionary — it is a deterministic interpreter of the user's pre-authorized rules. The discretion was exercised by the user when they wrote and confirmed the strategy.

Implication for attorney review: Confirm that this architecture is accurately described and that there is no step in the execution path where Raxx's system exercises judgment about whether to fire an order beyond the user's stated rules. If Raxx ever adds any adaptive or contextual execution logic (e.g., "skip this order if volatility is extreme"), that would change the analysis and require a separate attorney review.

Principle 4 — Retrospective only; never forward-looking

Statement: Raxx never predicts future performance. All AI-surfaced observations describe what has happened to the user on their own past trades. The platform uses historical markers (earnings dates, news events, volatility regimes) as context for past performance. No output on the platform says "this strategy will return X%," "this setup is likely to succeed," or any forward-looking framing.

Rationale: A notification framed as "your last 32 iron condor entries returned 88.4% — current conditions match those parameters" is a retrospective fact pattern. A notification framed as "this setup has 88% probability of success" is a forward- looking prediction. The latter is closer to "advice as to the advisability of" investing under the statute. Raxx's copy policy maintains the former framing strictly.

Implication for attorney review: Confirm that the distinction between historical fact framing and probability/forecast framing is legally meaningful under the § 80b-2(a)(11) "advisability" prong. Confirm whether the phrase "setup matches your 90-day winning profile" (current draft notification language) is defensible historical framing or whether the word "winning" introduces an evaluative judgment that crosses the line. Provide specific rewrite guidance on any copy that blurs the retrospective/forward boundary.

Principle 5 — No fiduciary relationship

Statement: Raxx does not hold itself out as a fiduciary for its users. Raxx has no ongoing portfolio management relationship with users. Users retain complete decisional autonomy — they activate strategies, they can deactivate them at any time, and they can override or ignore any notification. Raxx's relationship with the user is a software-as-a-service subscription relationship, not an advisory or management relationship.

Rationale: The Investment Advisers Act's fiduciary duty attaches to registered investment advisers. The Lowe v. SEC (472 U.S. 181, 1985) analysis distinguishes between the "trust and reliance" relationship characteristic of an adviser and a publication relationship. Raxx's design — user retains control, no platform discretion, notifications are informational — is closer to the publication model on the "fiduciary relationship" dimension even though personalization makes the publisher exclusion itself unavailable.

Implication for attorney review: Confirm whether the TOS should include an explicit "Raxx is not your investment adviser; you are the decision-maker" user acknowledgment, and what specific language survives challenge. Confirm whether any feature of the MBT product (e.g., automated order execution without per-trade confirmation) creates a de facto fiduciary relationship regardless of how the service is labeled.

Principle 6 — The broker is the user's decision

Statement: Raxx's customer-facing copy never names the underlying broker infrastructure (Alpaca, SnapTrade). The user connects "their broker." Raxx is the strategy layer. The broker relationship is between the user and their chosen broker — Raxx is not a party to that relationship other than as a technology intermediary.

Rationale: Avoiding explicit broker endorsement reduces Reg BI indirect exposure (the theory that a platform "recommending" a specific broker to retail customers might face introducing-broker analysis). Raxx presents the default broker as a utility choice for new users, not a preferred or endorsed option.

Implication for attorney review: Confirm whether Raxx's presentation of Alpaca as a default broker for new users — without revenue-sharing and without explicit endorsement copy — is distinguishable from an "introducing broker" arrangement under Reg BI or FINRA rules. Confirm what changes, if any, would create introducing-broker exposure (e.g., adding a referral fee from Alpaca per account opened).

Principle 7 — No AUM, no commissions, no custody

Statement: Raxx's entire revenue model is subscription fees. No revenue is earned from trade execution, asset management, custody, or any broker-linked compensation. Raxx does not hold or have access to user funds. Raxx does not earn a fee based on how much money a user manages or how often they trade.

Rationale: The "compensation" prong of the SEC Release IA-1092 three-prong test is technically satisfied by a subscription fee even if the fee is not specifically for advice. However, the compensation prong only matters if the advisory element is also present. Where no advisory activity exists, the compensation source is irrelevant to the IA Act analysis. Keeping revenue purely subscription-based also avoids the "referral fee = introducing broker" exposure described in Principle 6.

Implication for attorney review: Confirm that a purely subscription-based revenue model, with no AUM fee, no trade-linked compensation, and no broker referral fees, is the strongest available revenue posture for avoiding the compensation-prong analysis. Identify any revenue structure change that would reopen the compensation-prong question or create additional registration exposure.

These principles reflect the operator's understanding of the product as of 2026-05-13 UTC. The attorney's review — not this document — is the authoritative analysis of their legal implications.