Moe: Daxos Teardown v3

Moe treats your possessions like a brokerage account. Point the camera at an item, a vision model identifies it, a valuation call with web search prices it, the item lands in a stock-style portfolio, and a daily 8am cron re-values everything tracked [C2 50:38]. One tap cross-lists to eBay and Depop through developer APIs and to Facebook Marketplace through a headless browser driving the user's own OAuth session [C2 10:40]. Pricing: 5 free scans, then $9.99/mo, plus 4% of external sales and 8% of native sales.

The app is real and independently verifiable: App Store id 6759468571, seller Market of Everything, Inc., released 2026-08-25, version 1.0 with no App Store version bump since, 191 MB, iOS 15.1+, 4.33 stars from 12 ratings (iTunes lookup, 2026-09-16). It is React Native [C2 11:15], so over-the-air JS updates would not show as App Store versions. Roadmap only, not shipped: item-for-item trading, public profiles, agents that message buyers, Android.

The pitch is that hyper-collectors and "depop girlies" have closets of idle inventory and no patience for eBay's listing flow, so Moe becomes the consumer-casual front end to every marketplace and eventually the exchange itself. The state of play is much earlier: 967 users on the founder's own 2026-09-11 dashboard, $175.00 of lifetime GMV across 5 orders, $20.60 of lifetime fee revenue, about 25 genuinely paying subscribers implied by $249.75 estimated MRR, and an LLM bill of $355.59 over 30 days that exceeds all revenue combined.

Quality signal from inside the room: Daxos partner Jake used the app and it "did hallucinate on a bunch of stuff... would say that it's a different thing than it is" [C2 12:30]. Joseph: not accurate on items, but the UX is "very fast," "pretty enjoyable" [C2 17:27]. Polished wrapper, mediocre recognition on the core function.

The cross-lister ceiling

Flyp is the whole argument. $15.8M raised, two technical ex-Mercari founders, four rounds, no priced round in over four years, and a homepage that now sells the product at zero. That is the revealed price of cross-listing. The companies that survived in the lane, List Perfectly, Crosslist and PrimeLister, never raised a dollar, and the biggest of them, Vendoo, took $125k of institutional money in its entire life after five years of paying reseller subscriptions. Moe charges $9.99/mo plus 4% on top of each marketplace's own take, into a lane where the best-capitalized player charges nothing.

What the incumbents ship free today

• eBay: Magical Listing generates title, description, item specifics, category, price and shipping from photos alone; 10M+ sellers had used AI features and 100M+ AI-assisted listings by April 2025 (digitalcommerce360.com/2025/04/15/ebay-ai-aided-tool-cut-steps-create-listings); Q4 2025 moved to agents that build the whole listing from photos (valueaddedresource.net/ebay-ai-magical-listing-revisited, 2026-09-15). Free; eBay monetizes an approximately 13.6% final value fee plus $0.30 to $0.40 per order (sellerfeecalc.com, 2026-09-15). Its February 2026 user agreement bans LLM-driven bots and end-to-end automated flows without human review (valueaddedresource.net, posted 2026-01-20).
• Facebook Marketplace: March 2026, AI draft listings from photos with a suggested price from local comps, shipping labels, AI auto-replies to buyers (pymnts.com, valueaddedresource.net hands-on, 2026-09-15). July 2026, a free standalone Seller app that builds title, description, price and category from one photo (thenextweb.com, techtimes.com, 2026-07-24). Nearly feature-for-feature Moe's scan-to-list flow, free, plus the sanctioned version of Moe's roadmap buyer-messaging agents. No public C2C listing API has ever existed; automation violates Meta's terms.
• Depop: one-photo AI listing since 2024-09-12 (news.depop.com); US seller fee cut to zero July 2024; inside eBay since the acquisition agreed 2026-02-18 (investors.ebayinc.com). Cross-listing tolerated under the 2026 terms if items are delisted promptly.
• Mercari: AI Listing Support since 2024-09-10 on GPT-4o mini (about.mercari.com); free; flat 10% seller fee; no public API, and an August 2024 change silently broke third-party delisting.
• Poshmark: Smart List AI since February 2025, Posh Lens visual search since 2023, free (digitalcommerce360.com, 2025-02-04); 20% fee over $15; suspended hundreds of cross-lister users in May and November 2025 (valueaddedresource.net).
• Vinted: no seller fees at all; US expansion January 2026 with roughly 800% download growth (appfigures.com, fashionunited.com, 2026-01-22).
• Google Lens: free point-camera price comparison at the camera and search layer since late 2024. Apple Visual Intelligence: free at the OS layer on iPhone 15 Pro and later, extended in iOS 26 (apple.com newsroom, 2025-06-09). Free-tier ChatGPT answers "what is this worth" from a photo.

Net: of Moe's paid bundle, only three pieces are not already free inside the platforms. The cross-marketplace layer itself (legal on eBay, gated on Depop, on Facebook only by terms violation), the portfolio framing (packaging, one feature flag from commoditization), and buyer-messaging agents (Meta shipped the sanctioned version; eBay contractually banned the unsanctioned one).

Verdict

Where Moe wins today, on the evidence: one axis against one subgroup. Against the pure cross-listers (Vendoo, List Perfectly, Crosslist, Flyp) it is better on valuation because they do not value items at all, and better on sticker price against the $29+ power-seller tiers. That is the whole list of wins, and both erode: a seller doing more than about $25/mo of external GMV pays Vendoo less at the $8.99 tier once Moe's 4% is counted, and Flyp is free. The valuation edge over the marketplaces' native flows is a number of unproven accuracy, contradicted by the one partner who tested it, delivered where eBay and Facebook already suggest prices from better data than Moe can reach. On gross margin per user Moe is worse than all twelve: it is the only product in the set whose dominant cost, the daily LLM revaluation of every tracked item, scales with free usage. On distribution and trust it is worse than all twelve by two to six orders of magnitude.

Who already ships the whole Moe loop: SellRaze is the near-complete clone with 39,616 ratings to Moe's 12, and the 2026-09-15 sweep adds Hero, Reclaim and Flippr doing the same bundle. What none of them does is the daily-revalued portfolio of arbitrary items, and the funded evidence says that piece only works inside a single vertical with liquid transaction comps: CollX raised $28.2M in cards, Collectr holds 64,973 ratings on $0 of recorded funding, and horizontal "value everything you own" has never raised past pre-seed anywhere in this data. The bundle's other unique element, one-tap Facebook Marketplace posting, is the illegitimate leg.

What a user loses by deleting Moe and using free incumbent features: the portfolio screen with daily revaluation, and one app instead of three. That is all. Valuation is free and unlimited through Lens, Visual Intelligence or ChatGPT against 5 scans then $9.99. Listing is free and native on eBay, Meta's Seller app, Depop, Mercari and Poshmark. Cross-listing at power-seller volume is $8.99 at Vendoo or $0 at Flyp. The user also stops paying a 4% override on top of each marketplace's own take and stops running terms-violating automation through their personal Facebook account.

The bar in ten lines

1. Median first institutional check for tools and valuation apps in this set is roughly $500k to $1.5M; the mode is YC-standard money of $125k to $500k. Only marketplaces and single-vertical card platforms opened above $3M.
2. The funded tools had traction Moe does not: CollX 600K users and 100M scans, SellRaze 200K users and 39,579 ratings, Vendoo five years of subscription revenue.
3. The marketplaces raised institutionally only after network proof: Whatnot with same-day livestream GMV, Vinted with millions of members after five unfunded years, Mercari with 1M listings in year one, OfferUp after two years grinding to Seattle liquidity.
4. The only pre-traction raise in the whole set, Poshmark in Feb 2011, went to a founder who had already sold a shopping company, flanked by two engineer cofounders.
5. Technical founders are near-universal. The one non-engineer founder in the winners column, Depop's Beckerman, had an incubator writing his product for a year.
6. The ceiling belongs to demand networks: Whatnot at $20B disclosed, Depop $1.4B and Poshmark $1.2B into incumbents, Mercari's IPO at $2.39B. None is a tool.
7. The cross-listing ceiling is Flyp: $15.8M in, four years without a priced round, product now free. No cross-lister has ever reached Series B.
8. The valuation-app ceiling is CollX at $28.2M and a $50M Harmonic-estimated Series A valuation, earned in one liquid vertical. Horizontal "value everything you own" has never raised past pre-seed in this data.
9. About a third of the roughly 30 venture-funded tool and valuation companies in the set are dead outright and over half are dead or pivoted. The corpses cluster in exactly Moe's two lanes.
10. What has actually bought a $10M+ entry valuation here: 600K users, a prior exit, or same-day GMV proof. Nothing in the record shows a $10M post going to a pre-traction solo founder in this category.

Moe's place on this table. Below every funded row, including the dead ones. Encore died with institutional backing and a team of three. SellHound died with $353k and a real prototype pipeline. The nearest living peer on the identical wedge has 3,300x Moe's rating count and 200x its user count, and was priced last September at $500k on a Harmonic-estimated $2.0M post. On a per-user basis the two asks are three orders of magnitude apart: about $10 per user at SellRaze, about $10,340 per user at Moe. On the team axis Moe sits below the graveyard rows, not just below the survivors, because Gone, Lovd, OneShop and SellHound all had engineering teams and still died.

(1) Market and competition: 1.5 / 8

The venture-scale outcomes in this market are all demand networks, never tools, and the tool lane has a hard ceiling in the same data (Flyp free after $15.8M; survivors on $0 to $125k). The graveyard sits in Moe's two lanes and is large. Every leg of the paid loop is already free inside the platforms Moe depends on. On this evidence the answer to the venture-scale question is no for a horizontal seller tool: the market pays zero for cross-listing, pays subscription-SaaS money only to bootstrapped operators, and reserves the billion-dollar outcomes for whoever owns the buyers.

(2) Product and wedge: 3.0 / 8

Measured against the 12 closest live competitors, exactly one thing in Moe is not available elsewhere: the daily revalued portfolio of arbitrary items combined in one app with cross-listing. The feature matrix from the 2026-09-15 sweep shows no other app holding all three of arbitrary-item AI valuation, multi-marketplace cross-listing and portfolio revaluation, and the cross-listers confirm it from the other side, since Vendoo, List Perfectly, Crosslist, PrimeLister and Flyp do not value items at all. The problem is what the unique part consists of. Where "your stuff is a portfolio" has ever retained, it was vertical with liquid transaction comps. Moe's daily cron prices arbitrary objects by LLM plus web search, and the one Daxos partner who used it found it "did hallucinate on a bunch of stuff" [C2 12:30], so breadth is bought with accuracy the category's funded apps do not have to give up. The other differentiating leg, Facebook Marketplace, is not a product advantage but a compliance exposure, and Meta shipped the sanctioned version of the same flow free in March and July 2026. Against the only competitor with the identical wedge the measured gap is not close: 39,616 iOS ratings from a 2023-09-11 release against Moe's 12 from 2026-08-25.

(3) Traction: 1.0 / 8

Verified: 967 users, about 25 real payers implied by $249.75 estimated MRR, $175 of lifetime GMV across 5 orders, $20.60 of lifetime fees, all read off the founder's own dashboard in the Loom of 2026-09-11, plus 12 iOS ratings unchanged through 2026-09-16. Funded peers in this exact market had two to three orders of magnitude more at their first institutional check: CollX 600,000 users and 100M scans, SellRaze 200K+ users and 39,579 ratings, Vendoo roughly five years of paying subscriptions. The marketplaces were stricter still. Moe's commerce proof is $175, which in a market whose entire thesis is that items actually sell is the number that matters, and 5 orders in three weeks cannot be extrapolated. What is real and worth the point: 967 signups in about two weeks with zero marketing spend, consistent across the Loom, both calls and the dashboard.

(4) Unit economics: 1.5 / 8

Moe's own LLM ledger shows $355.59 of spend over 30 days against roughly $270/mo of revenue, so cost is about 142% of revenue and gross margin is negative today. The shape is worse than the level: 1,004 valuation.backfill calls at $0.2342 each, $235.16, is 66% of all spend, and that is the daily 8am cron revaluing already-scanned inventory [C2 50:38]. The dominant cost line scales with items tracked by all users while revenue scales only with the 2.6% who pay and the sliver of tracked value that ever sells. No competitor in the set carries that structure. The cross-listing SaaS survivors are flat-fee software with no per-item metered inference, which is exactly how List Perfectly and Crosslist live on $0 raised, and the marketplaces earn a transaction take that rises with the same GMV that drives their costs. Moe charges $9.99/mo plus 4% stacked on top of those same platform fees, so a Moe-originated eBay sale costs the seller about 17.6% plus the subscription, and the customer Moe says it wants, the heavy scanning collector with thousands of items, is its single worst account. It is fixable with ordinary engineering (cache valuations by product, revalue on price-move triggers instead of daily, cheaper backfill models, meter scans), and nothing in either call suggests the founder treats it as a problem: he presents the cost dashboard as a strength while claiming costs are "dropping 20-30% day over day" [Loom 01:04] over a chart rising to $62.80 in a day in the same frame.

(5) Team: 1.0 / 8

Moe is a solo non-technical founder with no verifiable engineer. Harmonic's department split shows engineering 0, data 0, product 0 and design 0 against a recorded headcount of 12; no engineer has ever been named; the two named non-founder hires are both non-technical. A technical founder is close to universal among the funded companies in this market: Whatnot's Logan Head coded the livestream app himself in six weeks, Vinted's Justas Janauskas wrote the code, Mercari's Shintaro Yamada had already sold a games company to Zynga, OfferUp's Nick Huzar was a prior CTO, Flyp was founded by two ex-Mercari marketplace builders, SellRaze has CTO Tyler Ma, Beni has CTO Celine Mol, and Poshmark's non-engineer CEO was a repeat exited founder flanked by two engineer cofounders. The one non-engineer founder in the winners column had the H-FARM incubator writing his product for a year. Moe sits below the graveyard on this axis too, since Gone, Lovd, OneShop and SellHound all had engineering teams and still died, and the founder's claim to be "leading a lot of the other engineering pieces" [C2 06:10] rests on a finance BA, no engineering skills on his own LinkedIn and a GitHub account created 2025-12-31 with zero public repositories.

(6) Founder trust: 1.0 / 8

The inflation was refuted by his own documents inside five days: on 2026-09-10 he claimed 5,000 downloads, 3,500 users overnight and MRR "right around 10k" [C1 45:22, 50:35, 50:44]; on 2026-09-11 his own dashboard showed 967 users and $249.75; on 2026-09-15 he read "967 users, that was Friday" off the same dashboard without acknowledging the earlier number [C2 53:38]. He also sold a fabricated resume for a named third party, describing his data science lead as having "ran Data Science at DraftKings for the last five years and then three years of Nike" [C1 42:58] when the only matching person, Yagev Levi, is a May 2023 BU computer science graduate with about two years of IC time at DraftKings who works full time at another company in Tel Aviv (Apify LinkedIn pull, 2026-09-15). Two further items are undisclosed rather than inflated: he sued his prior employer and its US chief in federal court while citing that employer as pedigree (D. Colo. 1:24-cv-03084, judgment 2026-06-02), and he left two entities to lapse without winding them down. The background sweep found real clean ground, which is why this is 1.0 and not 0. Against the category this matters more than usual, because the funded comparables all sold verifiable counts (CollX's 600,000 users, SellRaze's 39,579 ratings) and the two claims that would most distinguish Moe from the graveyard, its user growth and its team, are precisely the two that failed verification.

Base

Mean of the six parts is 1.50. Base is taken at 1.75. The quarter point above the mean credits what independently verified: a real shipped iOS app with polished UX, 967 signups in about two weeks with no marketing spend, honest cost instrumentation, candid answers on product limits, and stable round terms across both calls. It stays below 2 because the three parts that decide whether a company in this market survives, the category's revealed economics, commerce traction, and whether a technical team exists, are all bottom-decile, and because the one structurally unique feature depends on a Meta terms violation whose blast radius is the user's own account. The 0.25 adjustment is explicit and reversible; a reader who declines it lands at Company 3.50.

Call 1, what he claimed

"3,500 users overnight yesterday," about 5,000 downloads, 200 to 300 users a day "all converting to pro" [45:22]; MRR "right around 10k" [50:44]; 70% download to subscription [50:33]; sell-through 15 to 20% in week one [45:35]; average item GMV $400 [46:29]; team of 7 "all based in the US" [09:16], every engineer 7+ years, average salary about $125k [52:50, 53:39]; DS/ML lead "ran Data Science at DraftKings for the last five years and then three years of Nike" [42:58]; eBay "signed agreements... they're done" [30:47]; "full end to end partnership with Meta currently right now" [29:59]; Mercado Libre signed, live early Q1 [32:53]; burn $20k/mo, 14 months runway; $700k of the $1M in.

Call 2, what changed

Users became "967... that was Friday... just approaching 2,000 now" [53:38], his own refutation of 5,000, delivered as if the first claim had never been made. MRR was never mentioned again. The 70% became "starting a subscription path," free trial or paid [33:49]: same number, new denominator, unacknowledged. GMV $400 became "hovered around the same mark, which I think was right, 300" [54:07]. The all-US team became engineers in Spain at EUR 50k and EUR 90k [05:47 to 06:04], which breaks both the $125k average and the 7-years-minimum claim. New extraordinary claims replaced the fallen ones: churn "sub two and a half percent" [34:04] on a one-week beta, and 97% trial to paid [41:25], irreconcilable with a $249.75 billing panel. The seed narrative escalated to a16z, Lerer Hippeau, Khosla, Headline, Bling and True "circling," "term sheets this week" [01:13], while conceding "we don't have a term sheet obviously yet" [03:41]. Ownership went "I own 100" to "95" to employees at 0.5 to 1% each to a pool "10% max" [07:12 to 08:34]; it does not add up, and no cap table exists.

How he handled pushback

Smoothly, never defensively, across 140 minutes without irritation. The technique is redefinition and authority-borrowing: the 70% survives by changing its base; Meta risk is answered with an unnamed advisor's blessing and a claimed partnership; fundraise doubt is answered with six VC names, two of them mangled ("Jason Horowitz," "Larrey Hippo"). Questions with no good answer (DAU, asked directly at [53:13], the MRR update, items actually sold) get adjacent answers. The one candid lane is product limits: video scanning "basically crashes the phone... within 30 seconds" [58:26], tap-to-scan is "basically static photos" dressed up [59:18], the scanner misidentifies, churn will rise toward 10% [41:58]. Credibility is bimodal: believable on product, not on metrics, partnerships or money.

Best quotes

• [C1 45:22] "We gained 3,500 users overnight yesterday and in the last week we've been growing about 200 to 300 users a day, all converting to pro subscriptions."
• [C1 50:35 to 50:44] "we've had about 5,000 downloads... our MRR right now is right around 10k."
• [C2 53:38] "We're at 967 users on... That was Friday, so four days ago. We're just approaching 2,000 now." His own refutation of both lines above.
• [C1 30:46] Mark: "They're like signed agreements?" Erik: "Yeah, they're done."
• [C1 09:16] "a team of seven people currently right now, all based in the US" against [C2 05:47] "there's a couple of engineers that are based in Spain."
• [C2 41:25] "97 of the people that go on to a free trial end up converting to being a paid."
• [C2 07:12 to 08:20] "I own 100... [Mark: what's your ownership?] 95."
• [Loom 01:35] "This basically shows our K factor, which we're hovering around 1.2, 1.4 right now", over a screen of company-minted single-use waitlist codes.
• [C2 58:26] "It basically crashes the phone... after I scanned more than four items and basically within 30 seconds shorts out the phone." The candid register he is capable of.
• [C2 63:19] Joseph: "I know you sent over the safe. We're all buttoned up on that. We can execute that like probably in the next day or so." The reason this teardown exists.

Erik Schutzler, founder and CEO, claims 95% ownership REAL, INFLATED ~2x, ADVERSE

Verified: real person (LinkedIn, Harmonic 94373779, App Store seller, side-letter signatory). BA Finance, University of Denver 2014 to 2018; no technical education; GitHub account created 2025-12-31 with zero public repositories. T-Mobile about 1.5 years; Amazon 2 years (his own LinkedIn: Middle East, UAE, Saudi, Egypt, "$30M revenue"); PingPong Payments about 2 years (US employee 30, then head of sales North America); Sport IQ, a dead DTC cream brand with no exit; Yiftee advisor. Almost certainly the son of Michael Schutzler (ex-CEO of Classmates.com and Livemocha, WTIA CEO), co-governor of Erik's college LLC at the same Seattle address (WA registry 604083592), which explains the Seattle angel network.

Claimed against found: "4.5 years early-stage Amazon plus about 2 years at an Amazon subsidiary in APAC" [C1 07:48] when his own LinkedIn says 2 years, and PingPong is a Hangzhou payments company, not an Amazon subsidiary. Claimed country launches include Japan (launched Oct 2000, when he was about 4) and Turkey (before he joined). "I'm leading a lot of the other engineering pieces" [C2 06:10] is not credible against a finance degree and an empty GitHub.

"Yagev" (Yagev Levi), claimed DS/ML lead RESUME FABRICATED

Erik: "ran Data Science at DraftKings for the last five years and then three years of Nike before that" [C1 42:58], with 3 engineers under him [44:41]. The only matching person (linkedin.com/in/yagevlevi, Apify 2026-09-15): BA Computer Science, Boston University, May 2023; about 2 years of IC ML-infrastructure work at DraftKings; about 1.5 years of pre-graduation Nike co-op; MS at Georgia Tech in progress; currently "AI @ terra" (Terra Security, Tel Aviv), started about a month ago; no Moe anywhere on the profile. The claim assigns 8 years of department-head seniority to someone with about 3.5 total who works full time elsewhere. False as delivered.

Janice Cheng, Founding Head of Growth REAL, CLAIM CONTRADICTED

Real (Harmonic 199300379, since July 2026). Claimed to have "built organic growth for three huge applications to number one in the App Store, including Hinge" [C1 43:49]. Her LinkedIn: Aavia plus fractional gigs (Dipsea, Rodeo, Ditto), no Hinge, no number-one apps. Likely fractional.

Jordan Novack, Head of Social REAL, POSSIBLY GONE

Sports social background (UFC, XFL, PitchBook), non-technical; the v1 LinkedIn pull showed the tenure end-dated 2026-09-01 while the Harmonic company record still lists the seat as current. Either way not an engineer.

Engineers, designer, advisors, investors ZERO VERIFIABLE

• Engineers: zero on Harmonic (headcount 12, engineering 0, data 0, product 0, design 0), zero on LinkedIn, never named. The all-US claim [C1 09:16] was false when made, by his own later statements.
• Design head: unnamed. "OfferUp founder" advisor: unnamed, and the "nearly a billion in revenue" attached to the claim is unsupported anywhere.
• Facebook Marketplace advisor: "one of the first creators" [C1 29:51] becomes "the first employee" [C2 44:47]. Mutating, unverifiable.
• Investors: an SPV, ex-marketplace angels, "a VC," the Moto Pizza founder's family office. Harmonic: funding_total $0, 0 rounds, investors []. EDGAR: no Form D.

Team verdict against the category

Every surviving company in all three sub-categories had at least one technical founder writing the product, and most had two or three. Moe has none, and the one hire whose technical credentials were sold to Daxos has a fabricated resume. That is a below-graveyard position on the axis this market treats as table stakes.

Verdict

Not a fraudster on the public record. Zero defendant-side lawsuits, no bankruptcy, no liens or UCC filings where searchable, nothing on OFAC, nothing at CFPB, BBB, Trustpilot or Ripoff Report, and an internally consistent identity (Eastside Catholic 2014, University of Denver 2014 to 2018, Seattle roots). One federal case ever, and he is the plaintiff with no counterclaim on the docket. The pattern of embellishment is confirmed, but the proof lives in his own pitch, not in court records.

The employer lawsuit

Schutzler v. PingPong Global Solutions, Inc. and Kenny Tsang, D. Colo. 1:24-cv-03084-RMR, filed 2024-11-04, employment discrimination, Erik the sole plaintiff, jury demanded. Defendants answered 2025-01-06 with no counterclaim on the docket, which matters: PingPong did not accuse him of anything. Then a long discovery grind with five stipulated-looking extensions, sealed filings in April and May 2026, an attorney withdrawal granted 2026-05-08, and judgment entered 2026-06-02 (entry 57). No motion to dismiss, no summary judgment, no trial. That pattern is most consistent with a settlement reduced to judgment or an accepted offer of judgment, but the prevailing party and terms need PACER (about $3). Suing a former employer is not misconduct. The issue is that he cited PingPong as pedigree across 140 minutes of calls and never mentioned he sued them and their US chief, and that the judgment landed three months before he pitched Daxos.

What Daxos must ask directly: the PingPong judgment and why it was never mentioned; where Market of Everything, Inc. is incorporated, with the certificate; the $700k, with the Form D or signed closing docs and wire receipts; DealsDepot L.L.C. and Schutzler Enterprises LLC, what they were and why abandoned rather than dissolved; what legal entity operated Sport IQ and who owns the still-live Amazon listing; whether any App Store launch reviews were solicited or compensated; whether Michael Schutzler is an investor, advisor or LP in the SPV.

The arithmetic, all from the founder-controlled dashboard

• Gross margin negative by design. Revenue about $249.75 MRR plus $20.60 of lifetime fees, roughly $270/mo, against $355.59/mo of LLM alone. LLM spend is 142% of MRR; margin about negative 32% before AWS, Steel sessions and Stripe.
• Revenue per user $249.75/967 = $0.26/mo. LLM cost per user $355.59/967 = $0.37/mo. The average user costs about 40% more than they pay.
• GMV per user $175/967 = $0.18 lifetime. Of $71,692 of tracked "inventory value," 0.24% ever transacted.
• Structural bleed: 66% of LLM spend is the daily revaluation cron [C2 50:43]. Cost scales with items tracked; revenue scales with the 2.6% who pay. The dream customer, the thousand-item closet, is the worst unit economics.
• Fixable in principle (cache by product, trigger-based revaluation, cheaper backfill models, metering), but he presents the cost dashboard as a strength while claiming a trend his own chart refutes.
• Category context: no competitor in the set carries this cost structure. The cross-listing survivors are flat-fee software with no per-item metered inference, which is exactly how List Perfectly and Crosslist live on $0 raised. The marketplaces earn a take that rises with the same GMV that drives their cost.

Instrument: $1M post-money SAFE, $10M post cap, 10% discount, MFN (verbal, C1 [49:12]; the memo adds a 1x cash-out floor). The SAFE was sent to Daxos after call 1 but is not in the data room, so nothing here verifies cap, discount, MFN, floor or governing law from a primary instrument. About $700k claimed committed, about $300k open.

Side letter (Pro Rata and Investor Rights Agreement, read in full): pro rata into the priced round; participation in any earlier or subsequent convertible up to the pro rata percentage; Major Investor designation on a commercially reasonable efforts standard; quarterly unaudited P&L, balance sheet and KPIs within 45 days plus annual financials and a quarterly cap table on request; a QSBS covenant; a capitalization rep (section 6 reps that a cap table was delivered, which it was not); free assignment to affiliates and SPVs. Genuinely good rights for an angel check. Not signed: both signature blocks blank, date "September ___, 2026," despite the _EXECUTION filename. No board seat, no observer, no vetoes, so the rep is the only enforcement lever.

Structure red flags

No Form D for a claimed $700k Reg D raise, which is late, unfiled, or evidence the money has not closed; each is a different problem. Entity unverified: no charter, no good standing, no state registration findable in NY or WA, Delaware captcha-walled. Cap table never delivered despite the section 6 rep. Employee equity is handshake-stage, since "I own 100" and a promised 10% pool cannot both be true unless the grants are unissued, which means no adopted plan, no vesting, no 83(b), and a pool still to be created that dilutes Daxos at the seed. The prior SAFE stack is opaque, so both the MFN check and the section 2 participation right are blind. IP is unassigned as far as the room shows: engineers in Spain, no PIIAs, no proof the company owns its code.

What to ask for before any wire (items 1 to 3 are the gate; a founder whose numbers are real produces them in a day): countersigned SAFE and side letter with terms read from the instrument; the cap table section 6 already promises plus every outstanding convertible and side letter; wire confirmations for the $700k and the Form D or counsel's written explanation of its absence; certificate of incorporation, good standing and board consents; equity plan, founder stock purchase agreement and 83(b) elections; signed IP assignments from every engineer and contractor including Spain; read-only Stripe or RevenueCat reconciled against the $10k MRR claim; the OfferUp advisor's name and agreement.

• Facebook Marketplace, the load-bearing leg. No public C2C listing API has ever existed. Moe's own description: "a headless browser spins up through an oauth token... and goes and actually posts it in Facebook Marketplace" [C2 10:40], plus in-account buyer-messaging agents [C1 18:40]. That is browser automation of user accounts against Meta's automated-means terms, whatever the "we're not scraping so we're fine" theory [C2 44:27]. The failure mode lands on the user's own Facebook account, the classic cross-lister death; Vendoo keeps a standing help article for users blocked from listing on Facebook (help.vendoo.co article 6817366, accessed 2026-09-16). The mitigation offered is an unnamed advisor's blessing plus a Meta partnership the memo says does not exist. Meanwhile Meta shipped the sanctioned version of the same flow free in March 2026 and a free standalone Seller app in July 2026.
• eBay. The user agreement effective 2026-02-20 bans LLM-driven bots and end-to-end automated flows without human review. The compliant lane is the official Sell API with human review, which is how Vendoo and List Perfectly survive; Moe's "direct API" is plausibly a self-serve developer key, not "signed agreements... done." The agentic buyer-messaging roadmap points directly at what eBay just banned. eBay is also the category's only likely acquirer, and it owns Depop.
• Depop. Terms prohibit bots and unauthorized API use; a gated Partner Selling API exists (partnerapi.depop.com, accessed 2026-09-16) and the 2026 terms tolerate third-party cross-listing with prompt delisting. "On their integration roadmap" is founder-claimed only [C1 30:33].
• Poshmark and Mercari, roadmap legs. Poshmark has the most documented enforcement against exactly this kind of automation, with suspension waves in May and November 2025 and listings mass-deleted. Mercari's August 2024 API change silently broke third-party delisting tools. Breakage is routine even for compliant incumbents.

1. Category economics. The tool lane's revealed price is zero (Flyp free after $15.8M), the survivors are bootstrapped, and no cross-lister has reached Series B. Horizontal "value everything you own" has never raised past pre-seed in this data.
2. Founder integrity. Every headline metric moved 3x to 40x in five days, never acknowledged; a named third party's resume was fabricated; his own Amazon record is doubled against his own LinkedIn; the PingPong suit went undisclosed while PingPong was cited as pedigree. Everything else in diligence becomes founder-gated.
3. Negative gross margin by design. $355.59 of LLM spend over 30 days against about $270/mo of revenue, 66% of it a daily cron that scales with inventory rather than sales. Growth makes it worse, and the best customer in the pitch is the worst account in the ledger.
4. No technical team. Solo non-technical founder "leading engineering," zero verifiable engineers, "own models" contradicted by a ledger literally named sonnet-vision, and code ownership open with no PIIAs.
5. Platform dependence is existential. The differentiating leg violates Meta's terms with users' accounts as collateral; eBay banned the agentic roadmap and owns two of the three rails; the acquirer of last resort ships the feature free.
6. Commodity wedge with better-executed competition. SellRaze ships the same bundle with 3,300x the review count at one twentieth the asked valuation, and Hero, Haz, Tradepost, Firesale, Minimist, Reclaim and Flippr attack the same wedge at lower prices.
7. Paper risk. Unsigned everything, no cap table, no Form D behind the claimed $700k, informal employee equity, entity unverified. The section 6 rep is the only lever and it reps an undelivered document.
8. Price. $10M post on 967 users and $175 of lifetime GMV, in a category whose live benchmark for the same product is a Harmonic-estimated $2.0M.

PASS. Company 3.75 / 10 (Base 1.75 + EV 1.00 + Geo 1.00 + Chain 0.00). FOR DAXOS 1.5 / 10. Do not execute the SAFE, at any check size, at this cap.

The honest bull case: a real shipped app with about 967 genuine signups in two weeks on zero marketing spend, a founder with real marketplace GTM history and unusual candor about product limits, cost telemetry most seed founders do not have, and a large secondhand backdrop with a live strategic acquirer. Why it still fails: the market it is in has produced venture-scale outcomes only for demand networks, priced cross-listing at zero, and buried about a third of its funded tool and valuation companies; the funded peers on the identical wedge carried 200K to 600K users at their first institutional check while Moe carries 967 and $175 of commerce; the team the premium rests on does not exist as described; the unit economics are negative by design and presented as a strength; and the load-bearing integration violates Meta's terms with users' accounts as collateral.

Exact conditions to reopen. All of these, not a subset.

1. Cap re-anchored to $2.5M post or lower, with the SAFE and every prior convertible produced showing cap, discount and MFN. At $10M the answer is no at any check size.
2. Proof of money in: wire confirmations or bank statements for the claimed $700k, the filed Form D, and a delivered cap table with the option pool sized in writing.
3. Corporate basics: certificate of incorporation and Delaware good standing, board consent authorizing the SAFE and the side letter, both signature blocks executed.
4. A named technical cofounder or first engineer who is verifiable and full time, with signed IP assignments from every engineer and contractor including Spain.
5. An on-record written correction of the Yagev Levi resume claim, plus disclosure of the PingPong judgment terms. Entry 57 of D. Colo. 1:24-cv-03084 is about $3 on PACER; Daxos should read it either way.
6. Read-only access to Stripe or RevenueCat showing paying subscribers excluding comped Pro-forever accounts, MRR, and 60 days of cohort retention, reconciled against the $10k MRR stated on call 1.
7. Gross margin positive at the unit level after the caching and revaluation-frequency fix, shown on the cost dashboard for a full month.
8. Commerce proof, not listing proof: at least $25,000 of cumulative completed GMV with the order count and take collected, which is the number this market actually pays for.
9. A legitimate path for the Facebook leg, meaning either a sanctioned integration or the product surviving without it, plus written confirmation that the eBay and Depop access complies with the February 2026 terms.

If items 1, 4, 6 and 8 all clear, the right posture is a small relationship check at the re-anchored cap, not a full position.

• Verso, 7.0, consumer, a Kalshi power-user interface. Far above Moe on every axis that file records. Not a close call.
• Fictura, 5.75 (Base 3.75 plus EV 1.00 plus Geo 1.00), a student consumer-adjacent SDK asking $100k at a $10M cap. Same cap as Moe and the same kind of overstatement problem, but Fictura's contradiction was a marketing claim about a prior app, while Moe's live metrics were refuted by its own dashboard inside five days and a named third party's resume was fabricated. Moe belongs below Fictura's base.
• TextSidekick, 5.5 company and 3.5 fit (Base about 3.5), solo founder, YC, self-claimed traction, wedge owned by better-funded incumbents. Structurally the closest shape to Moe in the file. It rates higher because its $500k raise and YC validation are real and its traction was unverified rather than self-refuted. Moe below it is consistent.
• Cob Foods, 4.5 IS, a portfolio consumer company with $1.44M of actual revenue and a deteriorating funnel. Moe's $175 of lifetime GMV places it far below on execution evidence, the axis the house rubric weights hardest.
• Pinned Social, reviewed 2026-08-12 at a $12M cap on a four-month-old iOS app with 467 US ratings at 4.9 stars, three first-time founders and no announced institutional money. It was treated as a lottery ticket priced as if it were not, and the recommendation was to insist on a side letter and a cap table before a $25k check. Moe asks a similar price on 12 ratings, no verifiable engineer and a documented self-refutation, so anything above Pinned Social's posture would be inconsistent.
• Haloo and Pear are portfolio consumer positions whose entries in the distribution file read "details TBD," so they support no usable comparison today.

Net: the 3.75 company score and 1.5 fit sit where the consumer distribution says they should, below TextSidekick and Fictura and far below Verso.

• No web page was re-fetched on 2026-09-16 for this rewrite. Every product, pricing, app-store, press and platform fact is carried forward with the URL and access date recorded in the 2026-09-15 and 2026-09-16 research files. Two Harmonic calls were run fresh on 2026-09-16 (themoe.app, sellraze.com) and both matched the recorded values.
• Harmonic's enrich payload returns an empty funding_rounds array for every company in this set, so exact seed-round amounts for any company with three or more rounds are unavailable. Only totals, last-round figures, investor lists and disclosed or estimated valuations are Harmonic-sourced. The $2.0M SellRaze, $625k Vendoo and OneShop, $6M Encore and $50M CollX valuations are Harmonic ESTIMATED, not disclosed.
• All Moe-side economics come from the founder's own dashboard captured in the 2026-09-11 Loom. Best available provenance, not audited. No Stripe or RevenueCat read-only access was ever provided.
• Moe's claimed $700k of commitments could not be verified in either direction: Harmonic shows $0 and 0 rounds, EDGAR shows no Form D, and no SPV or wire documents exist in the room.
• Delaware incorporation is unverified. Delaware ICIS is captcha-walled, OpenCorporates 403s, the WA SOS full name search returned a bot gate, and the NY registry returned no entity match.
• The PingPong judgment text (entry 57, D. Colo. 1:24-cv-03084) is not in RECAP and needs PACER; who prevailed and on what terms is unknown.
• Take rates were not captured for Facebook Marketplace's current shipped-order fee, CollX's internal marketplace take, SellRaze's take, and Vinted. Several fee pages are JS-walled to curl.
• Reddit and reseller-community sentiment on Moe and on the competing tools could not be pulled; droplet IPs are blocked and DuckDuckGo rate-limited. There is no community-sentiment input to any score here.
• The competitor sweep was iOS and Harmonic only. Google Play and Product Hunt were not swept, so an Android-only competitor would have been missed, and the keyword sweeps took the top 50 of 856 and 249 matches, so the long tail is not exhausted.
• OneShop's status is contested (Harmonic OUT_OF_BUSINESS against an app update dated 2026-09-12) and is carried as a zombie. Hero's 2025-11 round amount is undisclosed in Harmonic, so Hero is not used as a price benchmark. Collectibles.com funding is ambiguous across two Harmonic records. Alt's Harmonic founding date (2013-03) conflicts with press (2020) and is reported as recorded. Depop and Poshmark M&A dates in Harmonic differ from press close dates and are reported as recorded.
• The Base weighting is a judgment. The six parts are scored on one scale, the mean is 1.50, and the 0.25 upward adjustment to 1.75 is explicit and reversible. A reader who declines it lands at Company 3.50.