Alpaca Market Data Rights — Sentiment Pipeline (Wave A)

Status: research-only. This document does NOT constitute legal or tax advice. Before shipping any customer-facing surface that ingests, scores, or re-displays Alpaca-sourced data, consult a technology-licensing attorney with financial data-vendor experience. The Schwartz IP engagement scope (trademark / IP assignment) does not cover vendor data agreements; Schwartz IP may refer you to a specialist. Licensed jurisdictions relevant: federal (SEC / NMS plan rules), Delaware (Alpaca securities entity), Pennsylvania (operator entity).

Last updated: 2026-06-05 UTC. Sources retrieved: 2026-06-05 UTC. Verify freshness of Alpaca Terms PDF (metadata date 2021-08-31 — confirm current version with Alpaca).

Relationship to prior memo: docs/blr/2026-05-31-alpaca-display-rights-memo.md covers the News API / headline display question. This memo covers a distinct question: whether Alpaca market data bars (price OHLCV) ingested via the shared server-side Raxx account (ADR-0014 Surface 1) can be used as input to the sentiment pipeline and whether the resulting sentiment signals can be re-displayed to Raxx customers. Read both memos together; they are complementary, not duplicative.

TL;DR

Alpaca's public Terms PDF prohibits redistribution of "Services and Content" to others "for any commercial enterprise" without prior written consent, and the bluntest statement in any sourced document — Alpaca's November 2022 Support article — reads as a blanket ban: "you cannot redistribute Alpaca API data." Whether a FinBERT-derived sentiment score counted separately from the underlying bars is a contractual interpretation question the public documents do not resolve. The pass-through exchange licensor agreements (NASDAQ OMX Global Subscriber Agreement and NYSE Market Data Display Services Agreement, both incorporated by reference into Alpaca's terms) impose their own display-fee and qualified-subscriber obligations whenever a vendor re-displays data to external users. The net result: three distinct regulatory layers (Alpaca T&C, NASDAQ OMX, NYSE/CTA) all require written permission or a separate license tier before Raxx can display Alpaca-sourced data — or signals derived from it — to customers. Next step: written inquiry to Alpaca partnerships and, in parallel, to the pass-through licensor contacts Alpaca names in the subscriber agreements.

1. Architecture Context (from ADR-0014)

Raxx uses one server-side Alpaca account — the operator's Algo Trader Plus subscription — for all market data, across all Raxx customers (ADR-0014 "Surface 1"). No per-user OAuth is involved for market data. The subscriber to Alpaca is Raxx (MooseQuest LLC dba Raxx), not the individual Raxx customer.

The Wave A sentiment pipeline (cards #1383, #1386, #1387, #1401–#1406, #1488, #1390,

1381) would:

• (a) Ingest Alpaca bars (OHLCV, real-time and historical) via the shared server-side account — this step is already in production (Antlers charts, MBT order matching).
• (b) Score ingested data against FinBERT or similar — this step runs inside Reasonator, Raxx's isolated scoring service, with no Alpaca-sourced text visible to customers at that layer.
• (c) Re-display the sentiment-derived signal in customer-facing UI surfaces (Pro tier: retrospective sentiment panel; Pro+ tier: real-time decision-window chip).

Steps (a) and (b) are internal. Step (c) is the disputed question.

The BYOB dimension: Raxx supports Bring-Your-Own-Broker (ADR-0014, project_byob_hybrid_strategy). A Raxx customer may or may not hold their own Alpaca account. This matters because the NASDAQ OMX and NYSE subscriber agreements contain "qualified subscriber" obligations — if the end user displaying data is not an individually qualified subscriber, the re-displaying vendor (Raxx) must obtain a redistribution license and pay per-user display fees. A Raxx customer who does NOT have their own Alpaca account is definitionally not a qualified direct subscriber of NASDAQ/NYSE data.

2. Layer 1 — Alpaca Terms and Conditions

2.1 Governing document

• Alpaca Terms and Conditions PDF (2021-08-31 metadata creation date, confirmed at https://files.alpaca.markets/disclosures/library/TermsAndConditions.pdf).
• Incorporated subscriber agreements: NASDAQ OMX Global Subscriber Agreement and NYSE Market Data Display Services Agreement (see §3 and §4 below).
• NOTE: The prior memo (2026-05-31) notes that AW-1 — the signed Algo Trader Plus subscription agreement — is NOT in the repository. The signed agreement may contain display-rights language that supersedes or refines what is in the public Terms PDF. This is a load-bearing gap.

2.2 Key clauses (previously sourced and quoted in 2026-05-31 memo, cross-referenced here)

Personal and Non-Commercial Usage / User Application notice: The Terms PDF states that if you wish to use Services and Content "for any other purposes, including without limitation commercial usage, or making the Services and Content available to others through your own application (a 'User Application'), you shall provide Alpaca with 30 days advance written notice prior to making such User Application available to others."

Raxx is a User Application. The 30-day advance notice has not been sent as of 2026-06-05.

Content redistribution prohibition: "No part of the Service or Content may be copied, reproduced, republished, uploaded, posted, publicly displayed, encoded, translated, transmitted or distributed in any way (including 'mirroring') to any other computer, server, web site or other medium for publication or distribution or for any commercial enterprise, without Alpaca's express prior written consent."

Source: https://files.alpaca.markets/disclosures/library/TermsAndConditions.pdf (verbatim quotes sourced from prior BLR memo which retrieved and read the PDF in full; PDF binary not readable by this agent's WebFetch tool — see sourcing note at §8).

Redistribution support article (bluntest statement): "Unfortunately, you cannot redistribute Alpaca API data." Source: https://alpaca.markets/support/redistribute-alpaca-api (November 2022). Notably: this article does not distinguish news data from bars/OHLCV data. It covers "Alpaca API data" categorically.

2.3 The "Content" definition scope question

The critical threshold question for bars specifically: does "Content" in Alpaca's Terms include bars/OHLCV price data, or only news content?

Based on the Terms PDF structure and the redistribution support article's categorical language, the most defensible reading is that "Content" includes all data returned by Alpaca's API, including bars. The terms say "No part of the Service or Content," and "Service" encompasses all API endpoints. The redistribution article's categorical "you cannot redistribute Alpaca API data" (emphasis: API data, not news) reinforces this reading.

The operator-favorable argument — that bars are price facts in the public domain and therefore outside Alpaca's "Content" — faces two obstacles: (1) copyright analysis and contract analysis are separate; the contract may restrict what copyright law does not protect; (2) the pass-through exchange licensor agreements (NASDAQ OMX, NYSE CTA) independently impose redistribution controls on the same bars, regardless of what Alpaca's contract says.

2.4 The market data tier dimension

Alpaca's "About Market Data API" documentation confirms:

• Algo Trader Plus ($99/month): All US exchanges, OPRA options feed, unrestricted historical data, 10,000 API calls/min.
• Broker API tier: Separate pricing structure ($0–$2,000+/month), explicitly designed for platforms building products for end users.

Source: https://docs.alpaca.markets/us/docs/about-market-data-api

The prior memo (2026-05-31) identified this tier distinction as the single most load-bearing open question: Alpaca advertises that Broker API developers can build in-platform displays for end users — but this right is not automatically extended to Algo Trader Plus subscribers. The same analysis applies to bars: Algo Trader Plus is designed for a single trader's own use; Broker API is designed for platform builders re-displaying data to customers.

Alpaca terms confirm (from search results, June 2026): By selecting the Pro market data plan, users agree to the NASDAQ OMX Global Subscriber Agreement and the NYSE Market Data Display Services Agreement. This means both exchange licensor frameworks are directly in scope. Source: Alpaca disclosures page (https://alpaca.markets/disclosures).

3. Layer 2 — NASDAQ OMX Global Subscriber Agreement (UTP / CTA pass-through)

3.1 Governing document

• NASDAQ OMX Global Subscriber Agreement, incorporated by reference into Alpaca's Terms for Pro market data plan subscribers.
• Retrieved: https://files.alpaca.markets/disclosures/library/NASDAQ+OMX+Global+Subscriber+Agreement.pdf (PDF binary — not readable by WebFetch; see sourcing note at §8).
• Primary source cross-reference: https://gdpdatafeeds.nasdaqomx.com/GIWAgreement.aspx and https://www.nasdaqtrader.com/content/AdministrationSupport/AgreementsData/indemnifaddm.pdf

3.2 Key provisions (sourced from SEC filings, Finazon guidance, and Nasdaq Trader site)

Redistribution restriction: "Subscriber shall not redistribute any Information, in whole or in part, other than in accord with the provisions set forth in Section 1(b)." Source (paraphrase from search synthesis): https://go.factset.com/hubfs/Website_Downloads/Third-party%20Terms/nasdaq%20omx.pdf (PDF binary — verbatim from secondary source synthesis; confirm with primary).

Redistributor vs. subscriber distinction: Vendors (e.g., Alpaca) can be "Distributors" who provide APIs to clients without controlling the display. Where the Distributor does NOT administer an individual subscriber agreement with each end user, the Distributor must indemnify in the event of a claim. Source: Finazon SIP licensing guide — https://support.finazon.io/en/articles/7235634-a-practical-guide-of-signing-direct-agreements-with-utp-and-cta

Derived data definition (UTP rules, September 2023 update): "Derived Data consists of pricing data or other information that is created in whole or in part from the UTP Information. The Derived Data cannot be used to create other data that is recognized to be a reasonable facsimile for the UTP Information." Source: https://www.utpplan.com/DOC/datapolicies.pdf (UTP Data Policies, September 2023, effective January 1, 2024) — via Finazon synthesis.

Fee liability for derived data: Under UTP rules: AI/ML algorithms using individual ticker prices to create processed data ARE fee-liable. Aggregated metrics from multiple symbols (indices, portfolio valuations) are NOT fee-liable. Source: https://finazon.io/blog/understanding-sip-licensing-derived-data

Under CTA rules: "Derived data is always fee liable under CTA rules despite the intended use case." Source: same.

Non-display derived data fees: $3,500/month per use case (non-display derived data, CTA). Declared at onboarding. Source: https://finazon.io/blog/understanding-sip-licensing-derived-data

3.3 Application to the Raxx FinBERT pipeline

FinBERT sentiment scoring on Alpaca bars (OHLCV for individual tickers) is: - Per-ticker processing → falls into the UTP "fee-liable" AI/ML category. - CTA-derived data is always fee-liable regardless of use. - The "non-display derived data" category ($3,500/month) applies if Raxx uses the scores internally only (no customer-facing display). - If the scores ARE displayed to customers, the display-fee and qualified-subscriber obligations are triggered additionally.

The "reasonable facsimile" safe harbor: a FinBERT sentiment score (+0.74 positive) is NOT a reasonable facsimile for the raw OHLCV bars it was computed from. A score cannot be reverse-engineered to recreate the underlying price data. This satisfies the UTP "derived data" safe harbor on the technical definition — the score is legally derived data, not a redistribution of the underlying bars.

However: being classified as "derived data" under UTP/CTA rules does not mean "free from obligations." It means the obligations shift from redistribution-display fees to non-display derived data fees (and display fees on top if the scores are shown to customers). The sentiment pipeline likely generates both categories of liability.

4. Layer 3 — NYSE Market Data Display Services Agreement (CTA pass-through)

4.1 Governing document

• NYSE Market Data Display Services Agreement, incorporated by reference into Alpaca's Terms for Pro market data plan subscribers.
• Retrieved: https://files.alpaca.markets/disclosures/library/NYSE+Market+Data+Display+Services+Agreement.pdf (PDF binary — not readable by WebFetch; see sourcing note at §8).
• Primary source cross-reference: https://www.nyse.com/publicdocs/nyse/data/NYSE_Market_Data_Complete_Policy_Package.pdf and https://www.ctaplan.com/policy

4.2 Key provisions (sourced from NYSE public policy documents and search synthesis)

External Display License: An External Display License allows a customer to provide "Controlled Displays" to External Users. An External Use License permits a customer to provide a NYSE Data Product to External Users. These are separately licensed rights — not included in a basic subscriber status. Source: NYSE Market Data Policy Package search synthesis, https://www.nyse.com/publicdocs/nyse/data/NYSE_Market_Data_Complete_Policy_Package.pdf

Non-professional subscriber qualification: Vendors must qualify each end user as nonprofessional before they can access data at the nonprofessional rate. Incorrect qualification results in the vendor being liable for retroactive fees at the professional rate. Source: NYSE Market Data Policy search synthesis.

Non-professional display fee (real-time data): The CTA/UTP schedule references per-user fees for non-professional display of SIP data. For context, OPRA (options) charges $1.25/month per nonprofessional user for real-time display; CTA/NYSE equities have analogous structures (exact current rates: confirm with CTA plan at https://www.ctaplan.com/pricing — unsourced in this research pass).

Redistribution: "Except as permitted in the applicable subscriber agreements governing use of market data, there will be no redistribution of the data (including electronic) to clients, other organizations, or any person outside of the Customer's/Subscriber's organization." Source: NYSE redistribution synthesis from CTA Exhibit A policy documents.

Derived data and non-display fees: Derived data creation from NYSE data incurs non-display fees (e.g., for indices, financial products). The fee schedule is maintained at: https://www.nyse.com/publicdocs/nyse/data/NYSE_Market_Data_Pricing.pdf (May 14, 2026 version available — unsourced in detail by this research pass; confirm current rates).

4.3 Application to the Raxx situation

Raxx as an operator, using one shared Alpaca account for all customers: - Is a "vendor" in NYSE/CTA terminology — an entity that receives market data and re-distributes or displays it to external users (Raxx customers). - Requires an External Display License to provide Controlled Displays to its customers. - Must qualify each Raxx customer as professional or nonprofessional and report monthly. - Must have each Raxx customer sign a qualified subscriber agreement (or ensure the vendor-administered agreement covers them) before displaying real-time SIP data.

This framework applies regardless of whether the display is raw bars or sentiment scores derived from bars. The display of derived data that is traceable to SIP- sourced inputs likely requires the same licensing and qualification steps as the display of raw SIP data. The "non-display derived data" category provides a separate compliance pathway for backend pipeline use only (no customer-facing display).

5. Layer 4 — OPRA (Options Price Reporting Authority)

5.1 Scope

Relevant only if the sentiment pipeline ingests options data (Greeks, implied volatility, options bars). Reasonator's current design (per docs/architecture/reasonator/design.md) does not specify options data as a sentiment input; the design focuses on news headlines (News API) and equity bars. However, the Algo Trader Plus subscription includes OPRA options data, and the Reasonator API contract does not explicitly exclude options data.

5.2 Key provisions

Redistribution fee: OPRA charges a flat $1,500/month to any entity that re-displays OPRA data to external users ("whether you have one user or one million"). Source: https://www.marketdata.app/education/options/opra-fees/

Per-user display fees (real-time): - Nonprofessional: $1.25/month/user (volume discounts available). - Professional: $31.50/month/user or per device (no discounts). Source: same.

Exemption for delayed/historical data: If only delayed (>15 minutes) or historical options data is displayed, OPRA does not charge per-user display fees. Only the $1,500/month redistributor fee applies. Source: same.

User documentation requirement: Every user receiving real-time OPRA data through a platform must sign a formal OPRA Subscriber Agreement. The platform must store signed agreements and related user records for the user's membership plus at least three years. Source: same.

5.3 Application to Wave A

If the Wave A pipeline excludes options data (sentiment based on equity bars only), OPRA obligations are not triggered. If options data is included, the $1,500/month redistribution fee and per-user qualification apply. The Reasonator design should explicitly state whether options bars are in scope for sentiment scoring.

6. Three Plausible Answers (Likely-Yes / Likely-No / Depends-on-Specifics)

Answer A: Likely YES (operator-favorable interpretation)

Premise: A FinBERT sentiment score is Raxx's own work product. The bars are factual price data not protectable by copyright (Feist Publications, Inc. v. Rural Telephone Service Co., 499 U.S. 340 (1991), establishing that factual data lacks copyright protection). The score is a mathematical transformation that cannot be reverse-engineered to recreate the underlying bars. Raxx's model, Raxx's inference pass, Raxx's output. Alpaca's Terms restrict "the Content," and an independently computed score is not "the Content."

Conditions for "Likely YES" to hold: 1. Alpaca explicitly confirms in writing that sentiment scores derived from bars are Raxx's own work product, not "Content" subject to redistribution restrictions. 2. Raxx obtains separate redistribution/display licenses from NASDAQ OMX and NYSE (or confirms with Alpaca that their Distributor agreements cover downstream display by platforms like Raxx). 3. Raxx qualifies each customer as professional or nonprofessional and pays applicable display fees, or uses a delayed-data stream for the sentiment inputs (15+ minutes delayed data triggers significantly lower or no per-user display fees). 4. Raxx sends the 30-day advance written notice to Alpaca before any customer-facing sentiment display surface goes live.

Assessment: This path is legally achievable. The contractual work required is substantial but not unusual for a B2C financial data platform. The question is timing (pre-launch vs. post-launch) and which Alpaca tier supports it.

Answer B: Likely NO (conservative interpretation)

Premise: Alpaca's Terms do not distinguish "derived analytical score" from "the Content." The November 2022 support article's blanket "you cannot redistribute Alpaca API data" applies without exception. The UTP/CTA framework explicitly identifies AI/ML processing of individual ticker prices as fee-liable derived data. Without a specific written carve-out from Alpaca and without signed exchange licensor agreements, displaying any signal causally dependent on Alpaca bars to Raxx customers is not permitted under the current Algo Trader Plus subscription.

Assessment: This path does not require the sentiment pipeline to be abandoned — it requires an input-source substitution. If Wave A sentiment runs on a separately- licensed news feed (e.g., a direct Benzinga or MarketAxess agreement) rather than on Alpaca bars, the Alpaca/NYSE/NASDAQ/OPRA constraints are sidestepped. The engineering architecture (Reasonator, FinBERT) remains intact.

Answer C: Depends on Specifics (most likely near-term answer)

Premise: The documents create ambiguity at three specific junctures that only written Alpaca confirmation can resolve:

1. Does "Content" in the Alpaca Terms include bars/OHLCV, or only news/text data? The redistribution article says "API data" (bars included); the prior memo focused on news. This needs explicit Alpaca confirmation.

2. Does Algo Trader Plus carry any display rights for re-surfacing data to customers, or only Broker API? The docs advertise Broker API for this use case. Whether a written notice under the User Application clause gets Raxx the same right on Algo Trader Plus is unknown.

3. Do the NASDAQ OMX and NYSE agreements flow through to Raxx automatically (via Alpaca's Distributor agreements with the exchanges), or does Raxx need to sign separate subscriber agreements with each exchange? This is the most operationally consequential unknown — if Raxx must sign direct exchange agreements, the timeline for Wave A display surfaces extends significantly.

Assessment: This is the most realistic near-term framing. Answer A or B becomes determinative only after Alpaca responds to the outreach and after exchange licensor positions are confirmed. Wave A internal pipeline (scoring, storing) can proceed. Customer-facing display should remain feature-flagged off pending confirmation.

7. Architectural Implications

If Answer A (yes, with proper compliance):

No architectural changes needed. Compliance overhead only: - Send 30-day advance notice to Alpaca (per User Application clause). - Obtain written Alpaca confirmation that sentiment scores on bars are permissible. - Confirm exchange licensor pass-through coverage (or sign direct agreements). - Implement customer qualification workflow (professional vs. nonprofessional subscriber status capture, stored agreements, monthly reporting). - If real-time bars are the scoring input: per-user display fee reporting to NASDAQ/NYSE. If 15-min-delayed bars are the scoring input: reduced fee exposure. - For options sentiment (if in scope): OPRA redistributor agreement ($1,500/month) + per-user fees.

Estimated compliance timeline: 6–12 weeks to get written confirmations and sign exchange agreements, assuming no blockers. Wave A internal scoring can ship now; customer-facing display gates on signed agreements.

If Answer B (no, input-source substitution required):

Architectural change: replace Alpaca bars as the sentiment input with a separately- licensed data source. Options:

If Answer B, the cleanest engineering path is: Reasonator's input source becomes configurable. The scoring pipeline is source-agnostic (FinBERT takes text; bars feed is a config-driven provider). This architectural loosening has independent value for data-vendor diversification.

If Answer C (wait for confirmation, flag off):

No architectural changes required now. Both paths above remain open. The feature flag rack_enabled (per Reasonator design doc §7) already gates all customer-facing display. Internal scoring pipeline (batch scoring, audit log, sentiment_score_audit table) can proceed behind the flag.

8. Required Disclosures / Disclaimers (if Answer A)

If Alpaca and exchange licensor confirmations are obtained, the following disclosures are required or strongly recommended:

1. Subscriber qualification disclosure: At signup or at first access of a sentiment-enabled surface, Raxx must present and retain the exchange subscriber agreement (professional vs. nonprofessional qualification) for each user. Customer-facing copy: "By using this feature, you confirm you are a non-professional subscriber as defined by NYSE and NASDAQ rules" (or equivalent). This is a regulatory obligation, not optional.

2. No forward-looking framing (invariant I-1): All sentiment display surfaces must use retrospective language only. "Sentiment was X during this period" not "Sentiment is X, therefore consider Y." Per feedback_no_forward_looking_framing.

3. No broker naming (invariant I-7): "Market data sourced from licensed providers" not "powered by Alpaca" or "data from NYSE" (unless required by licensor). Per feedback_no_backend_branding.

4. "Powered by Alpaca" / Marks: The Terms prohibit using Alpaca's Marks without written consent. If a "Powered by Alpaca" badge is not wanted, no badge needed. If wanted, get written consent separately.

5. Data accuracy / warranty disclaimer: Alpaca's public disclosures state "Market prices, data and other information available through Alpaca are not warranted as to completeness or accuracy." Raxx's sentiment display surfaces should carry a commensurate disclaimer. (Reasonator design already accounts for this via invariant I-1 and the "Facts, not opinions" tagline.)

9. Wave A Card Impact Analysis

Estimate if Answer A confirmed: Wave A display surfaces unblocked in 6–12 weeks (Alpaca response time + exchange licensor agreement signing). Internal scoring pipeline (#1381, #1401–#1406) can ship in the interim.

Estimate if Answer B required: Wave A display surfaces unblocked in 10–20 weeks (alternative data vendor evaluation + contract + integration). Internal Reasonator architecture continues; input source changes.

10. Questions for Your Attorney (Technology-Licensing Specialist)

The Schwartz IP engagement (trademark / IP assignment) does not cover this domain. The recommended specialist type is a technology-licensing attorney with financial data-vendor experience. Securities counsel (if already engaged per docs/business-legal/securities-attorney-engagement-2026-05-13/) may also have relevant expertise.

1. Does a FinBERT sentiment score derived from OHLCV bars qualify as "derived data" under the UTP September 2023 Data Policies, and does that classification create any safe harbor from Alpaca's redistribution prohibition?

2. Is displaying a numerical sentiment score (no raw bars shown, no Alpaca-sourced text shown) to an authenticated paying customer a "redistribution" under Alpaca's Terms, the NASDAQ OMX Subscriber Agreement, and the NYSE Market Data Display Services Agreement?

3. Does Alpaca's Distributor relationship with NASDAQ and NYSE flow through to Raxx (meaning Raxx's customers are covered by Alpaca's exchange agreements), or does Raxx need to sign direct subscriber agreements with NASDAQ OMX and NYSE?

4. What is the correct per-user fee calculation for nonprofessional subscriber display of sentiment signals derived from CTA (NYSE-administered) and UTP (NASDAQ-administered) SIP data?

5. Does the 15-minute delay threshold (feeding delayed bars rather than real-time bars into Reasonator) materially reduce or eliminate per-user display fee obligations?

6. If Raxx upgrades to Alpaca Broker API, does that tier include display rights for building customer-facing data surfaces, and what compliance overhead does that introduce?

7. What is the minimum viable subscriber qualification workflow (professional vs. nonprofessional) for a SaaS platform whose customers are non-professional retail traders? Is a click-through attestation sufficient, or is a wet-signed agreement required?

8. The Wave A sentiment pipeline uses one shared Alpaca account for all Raxx customers (not per-user OAuth). Does this pooled-access architecture alter the exchange licensor analysis in any way?

11. Relationship to Prior Memo

The 2026-05-31 memo (docs/blr/2026-05-31-alpaca-display-rights-memo.md) covers: - Alpaca News API (headline text display to subscribers) — Scenario A - Sentiment scores derived from Alpaca news headlines — Scenario B - Attribution / Marks — Scenario C - Per-user vs. aggregate — Scenario D

This memo covers: - Sentiment scores derived from Alpaca market data bars (OHLCV) — new question - NYSE/CTA and NASDAQ OMX/UTP pass-through licensor obligations — new analysis - OPRA options data dimension — new question - Wave A card-by-card impact — new analysis

The core contractual tension is the same across both memos: the User Application notice requirement, the redistribution prohibition, and the tier question (Algo Trader Plus vs. Broker API). The exchange licensor dimension adds new compliance obligations that apply to bars display even if Alpaca grants a written carve-out for news sentiment.

Open items AW-1 through AW-5 from the prior memo remain open and are load-bearing for this memo as well.

12. Sourcing Notes

The following documents returned binary PDF content that could not be read directly by this research pass. Key clauses were obtained via secondary sources (search synthesis, Finazon guide, Hawk Alpha documentation) and cross-referenced against multiple independent sources. An attorney reviewing this memo should verify the primary PDF text directly.

13. Sources

This memo is research material only. It does not constitute legal advice. Before shipping any customer-facing Wave A surface that displays Alpaca-sourced data or signals derived from it, consult a technology-licensing attorney experienced in financial data vendor agreements and NMS plan (CTA/UTP/OPRA) compliance. The specific professional needed: a technology-licensing or financial-services regulatory attorney, NOT a trademark attorney. Schwartz IP (trademark scope) may refer you to a suitable specialist.