My Things
An automated system sourcing free items from Craigslist, Gumtree and Facebook Marketplace, then reselling them elsewhere.
Key outcome
Killed. Not scalable without escrow or secure payment infrastructure to reduce transaction risk.
02 / The problem
Thousands of usable items are given away for free every day on classified sites, while the same items sell for real money on other marketplaces. The gap is time and effort.
03 / The hypothesis
“If an automated system finds, scores and claims free listings, then relists them on paid marketplaces, the spread will be profitable at volume, because sourcing cost is near zero.”
04 / The solution
- →A monitoring layer watches free-item listings across classifieds.
- →Each listing is scored on resale value, distance and condition.
- →Promising items are claimed, collected, and relisted on resale marketplaces with generated descriptions.
05 / Build components
Watches free sections across sources.
Estimates resale value vs. effort.
Messages to givers to secure items.
[Add tool]
Generated titles, descriptions and pricing.
Identified as the missing piece.
06 / AI & LLM stack
- Build platform
- Lovable
- LLM(s)
- [Add model]
- AI use case
- [Add use case — e.g. deal scoring, listing generation]
- Automation
- [Add tools]
- APIs / integrations
- [Add tools]
- Hosting / infrastructure
- [Add tools]
07 / Distribution & GTM
- →Supply: free listings on Craigslist, Gumtree, Facebook Marketplace.
- →Demand: relisting on secondary marketplaces.
- →[Add channels tested]
08 / KPIs & results
Visitors
TBC
Signups
TBC
Paid conversions
TBC
Revenue
TBC
CAC
TBC
LTV
TBC
Conversion rate
TBC
Gross margin
TBC
Payback period
TBC
Retention / repeat
TBC
Items flipped
TBC
Values marked TBC are pending real data.
09 / Scale-or-kill
Kill.
Without escrow or secure payment infrastructure, transaction risk grew with volume. The model couldn't scale safely.
What worked
- +Free supply is abundant and easy to detect.
- +Automated scoring and listing generation cut manual effort.
What didn’t
- −Buyer payments and handovers carried fraud and no-show risk.
- −Logistics remained manual and didn't scale with automation.
Next time I’d test
- →Start on platforms with built-in payment protection.
- →Test a single high-value category first to improve margin per trip.
10 / Total cost & credits
- Lovable credits
- TBC
- AI / LLM credits
- TBC
- Hosting
- TBC
- Domain
- TBC
- Paid ads
- TBC
- Tools & subscriptions
- TBC
- Contractor / creative
- TBC
- Total experiment cost
- TBC
11 / Lessons learned
- 1.Automation scales software, not logistics.
- 2.Trust infrastructure is a prerequisite in peer-to-peer commerce.
- 3.Near-zero sourcing cost doesn't mean near-zero cost per sale.
12 / Next experiment
A narrower bet: tooling for professional resellers (sourcing alerts and listing automation) rather than doing the arbitrage directly.