Resollo

Hey IH 👋

I've been building Resollo for the past year — it's a marketplace where you upload photos of something you want to sell, and the AI writes the entire listing for you (title, description, category, attributes, condition, price).

The UX goal was: feels like posting a photo, not filling a spreadsheet.

AI autofill uses purchasable credits; manual listing is always available at no extra cost.

Tech stack: Next.js 16, Supabase (auth + DB + storage), Stripe Connect, OpenAI, Vercel.

Interesting architecture choices:

• AI autofill runs client-side image upload → signed URL → server-side AI vision call → structured JSON response

• Founding Member priority is implemented as a tiebreaker in the Postgres ORDER BY — no separate ranking table needed

• 8 languages via next-intl with per-route static generation

• Every active listing has its own dynamically-generated llms.txt (per the llmstxt.org standard) — not just specs, but a semantic profile (use cases, audience, key strengths). To my knowledge no other marketplace does this per listing, at scale; the index is a single DB query, TTL scales with activity

• The platform serves dynamic llms.txt (per-listing AI-readable files) and llms-full.txt (a platform-wide index) — both generated at request time from live DB data, never cached

Happy to answer any technical or business questions!

→ https://resollo.com/?utm_source=indiehackers&utm_medium=referral&utm_campaign=directory