SellPilot

What the company appears to be: SellPilot is a self-reported AI-powered tool for sellers of used goods. It processes product photos and optional seller context to generate structured selling guidance — including identification, condition, pricing scenarios, listing copy, and buyer responses. The system separates facts from uncertain inferences and provides actionable next steps.

What changed: The project was submitted as part of the OpenAI 2026 hackathon. It is described as a prototype built during a "Build Week" with no revenue or customer data reported.

Single most important open question: Is there evidence that the tool has been used by real sellers in real marketplaces, or that it has moved beyond a demo?

The description states that SellPilot processes one to four product photos and optional seller context. It claims to produce:

• Probable product identification
• Visible condition
• Missing details
• A recommended next photo
• Three clearly labeled price scenarios
• Listing copy
• Quick buyer responses

It does not claim to read live marketplace data, verify authenticity, or publish automatically.

The tool uses GPT-5.6 for multimodal reasoning and integrates with OpenAI's API via a server-side endpoint built with Next.js, React, and TypeScript.

Evidence: The author states this is what the product does.

Inference: The system appears to be designed to reduce uncertainty in the selling process by structuring AI outputs with provenance and confidence levels.

The tagline states: “Tu copiloto de IA para vender mejor: convierte fotos en análisis verificables, precios orientativos y anuncios listos para revisar.” (Your AI co-pilot for better selling: turns photos into verifiable analysis, rough pricing, and ready-to-review ads.)

The author's write-up claims that existing AI tools often hide uncertainty and make confident claims without showing evidence. SellPilot is positioned as a tool that separates seller-provided facts from AI-generated inferences.

Evidence: The tagline and the author’s own description reflect this positioning.

Inference: The product is trying to differentiate itself by transparency, not just automation.

The project description does not name specific customer segments or personas. It implies a target of sellers of used goods — likely individuals or small businesses selling items on platforms like Facebook Marketplace, eBay, or MercadoLibre.

Evidence: The author describes the problem as “selling a used item looks simple until the seller has to identify it, describe its condition, choose a price, write a trustworthy listing, and answer buyers.”

Inference: The tool is aimed at sellers who struggle with creating compelling, accurate listings — especially those without experience or tools to do so.

There is no evidence of pricing, monetization, or business model in the description. The product is described as a demo and does not claim to be live or generating revenue.

Evidence: The description states that the demo works without an account, and no pricing information is provided.

Inference: If this tool were commercialized, it might be sold via subscription or usage-based models, but there is no evidence of such plans.

The system uses:

• Next.js, React, and TypeScript for frontend
• Server-side endpoint calling OpenAI’s GPT-5.6 API
• Structured Outputs validated against a JSON schema
• High image detail input to the model
• Strict output formatting and provenance tracking
• Browser-based storage for drafts (for this Build Week version)

The tool is described as having a resilient fallback when OpenAI is unavailable, with simulated outputs explicitly labeled.

Evidence: The author states how it was built and what technologies were used.

Inference: The architecture suggests a lightweight, demo-grade system that could be scaled or integrated into larger platforms.

The project is described as a prototype built during a hackathon. It has no reported customers, revenue, or usage metrics. The public demo works without an account and stores data only in the browser.

Evidence: The description states it was submitted to a hackathon and that inventory and drafts remain in the current browser for this version.

Inference: There is no evidence of product-market fit or commercial traction.

The description does not mention competitors. It positions SellPilot as solving a gap in existing AI tools — those that “hide uncertainty” and make confident claims without showing evidence.

Evidence: The author states this is the problem it addresses.

Inference: Competitors likely include general-purpose AI writing tools, marketplace listing generators, or seller support platforms, but no names are given.

• No real-world usage: The tool is described as a demo with no evidence of adoption.
• Unproven commercial viability: No pricing, monetization or customer data.
• Limited scope: It does not integrate with actual marketplaces or automate publishing.
• Unclear scalability: The current version stores data in the browser and lacks durable storage.
• Dependency on GPT-5.6: If this model is not available or changes, the tool may fail.

Evidence: The description makes no claims about real-world use or business success.

1. Has SellPilot been tested with actual sellers in real marketplaces?
2. What is the plan for monetization and pricing?
3. How does the tool handle edge cases, such as low-quality images or ambiguous product types?
4. Are there plans to integrate with specific marketplace APIs or platforms?
5. What are the long-term technical dependencies beyond GPT-5.6?