Tutus, in five years: Listing your first item

A heavy Woolrich parka sits on a kitchen table in Lyon. On most peer-to-peer marketplaces in 2025, turning that coat into liquidity requires eight to ten minutes of manual labor: taking six photos under uneven lighting, measuring armpit-to-armpit spans with a tailor's tape, drafting a paragraph describing minor cuff fraying, and guessing a listing price based on erratic search results.

That manual bottleneck explains why millions of wearable items remain trapped in wardrobes across Europe. The future of re-commerce relies not on convincing users to spend more time filling out drop-down menus, but on eliminating manual input entirely.

Here is our explicit product target for the Tutus universe five years out. We are clear about the status of this architecture: the baseline escrow and catalog infrastructure are live today; the fully automated computer-vision pipeline and multi-universe fulfillment routing described below represent an AMBITION targeted for full deployment by 2030.

The Three-Second Capture Workflow

In our target model, a user opens the WEVONE interface, selects the Tutus universe, and pans their smartphone camera over the garment for three seconds. They do not pick a category, type a brand name, or select a size from a drop-down menu.

Mia—acting as the platform's central intelligence framework—ingests thirty video frames per second. The vision models extract three distinct layers of telemetry:

1. Structural Geometry: Spatial depth mapping determines shoulder width, sleeve length, and torso drop down to the millimeter, removing the need for ambiguous subjective size tags like "Fits like an M."
2. Material and Condition Matrix: The system scans high-density texture maps to identify fabric weave (e.g., 24-ounce Melton wool), surface pilling, missing hardware, or collar discoloration.
3. Brand and Model Verification: Logo vector matching cross-references vintage tags against our internal historical garment database, isolating the specific production range (for instance, a late-1990s Made in USA line).

Within two seconds of video capture, the interface presents a completed card. The user reviews the output, taps confirm, and the listing enters the Tutus universe index. No typing required.

Under the Hood: Escrow, Ledger, and Mia's Context Memory

Automating the listing interface is meaningless if the underlying settlement system cannot handle condition ambiguity. This is where WEVONE's specific platform architecture replaces standard marketplace design.

When a item is captured, Mia's context memory logs the visual telemetry directly into the Tutus universe ledger. This digital record captures the exact surface state of the garment before shipping. Instead of relying on vague text descriptions during a dispute, the buyer's unboxing video is computationally compared against the listing's initial spatial map.

The transactional escrow mechanism adapts dynamically to this confidence score. For a seller with a high platform contribution score and an unambiguous visual capture, Mia reduces the standard escrow dispute window from 48 hours to 6 hours post-delivery, unlocking funds near-instantly. For items where material integrity shows higher variance—such as unbranded vintage or heavily worn leather—the escrow hold parameters automatically extend, holding funds in the neutral vault until the buyer accepts or the inspection window elapses.

Furthermore, pricing is not generated by arbitrary user guesses. Mia evaluates recent cleared transactions across European peer-to-peer platforms, adjusts for inflation and hyper-local seasonal demand in the seller's specific region, and proposes a dual-price band: a quick-sale target (estimated execution within 48 hours) and a fair-market target (estimated execution within 14 days).

Integrated Fulfillment Across Universes

Listing an item is only half the operational problem; getting the physical box from a closet in Lyon to a buyer in Antwerp represents the remaining friction. Traditional platforms rely entirely on third-party drop-off networks, forcing sellers to print labels, find cardboard boxes, and walk to drop-off points.

By 2030, Tutus will not operate in isolation. It relies on structural cross-pollination with WEVONE's Pilote (co-transport) and Mission (local micro-tasks) universes:

• Zero-Packaging Pickups: When a seller accepts a local purchase offer, Mia checks active routes in the Pilote universe. If a verified community member is already commuting between the seller's neighborhood and the buyer's district, the item is routed via peer-to-peer transport without requiring external shipping labels or single-use packaging.
• Localized Micro-Fulfillment: If a seller lists ten items simultaneously, a local service provider from the Mission universe can be dispatched to collect, inspect, and temporarily store the garments in a nearby neighborhood micro-hub, handling shipping on behalf of the seller for a fraction of the standard platform commission.

Honest Limitations and Open Engineering Questions

Vision documents often present idealized pathways while ignoring physical realities. We acknowledge several hard constraints in this architecture that remain unsolved in our current beta models:

• Tactile and Odor Telemetry: Computer vision can identify a surface stain or frayed hem, but it cannot detect lingering smoke, perfume, or fabric stiffness. Odor remains the leading cause of subjective buyer dissatisfaction in second-hand apparel. Until sensor technology advances, subjective scent claims must still be handled via human arbitration.
• Unbranded Vintage Edge Cases: For garments lacking tags, logos, or distinct historical silhouettes, Mia's automated valuation model defaults to broad material-weight assumptions. In testing, unbranded 1970s items exhibit a pricing variance of up to 40% against actual collector willingness-to-pay.
• Data Privacy Boundaries: Generating localized demand models requires spatial and behavioral data. Ensuring Mia processes visual capture solely for item analysis without ingesting background home environment data requires strict, client-side video frame masking.

The Shift from Manual Input to Algorithmic Trust

In 2023, Vinted reported over 500 million registered items, yet platform interaction patterns showed that listing drop-off rates remained highest during the manual photo-and-description phase. Sellers abandon draft listings when faced with repetitive text fields.

Our PROJECTION is that reducing listing time from eight minutes to under ten seconds will increase localized second-hand inventory density by a factor of four within targeted urban zones. By transforming listing from a manual data-entry task into an automated capture process backed by cryptographic ledger verification, Tutus removes the psychological and operational friction of re-commerce.

The goal is not simply to make selling clothes faster. The goal is to render the distinction between owning an unused garment and holding active liquid capital entirely invisible.