Etsy POD AI Analyzer
AI-driven POD opportunity discovery across regions
A 5-signal scoring engine that scans keywords across six Etsy regions and pairs them with self-hosted Ollama models to generate listing titles, tags, pricing strategies, and design briefs.
Open the live appproblem
The hard part of print-on-demand is not designing — it is finding a keyword with high demand, low competition, and a region where you can win. Doing this by hand for one keyword takes an evening; doing it for fifty takes a month.
solution
Next.js + FastAPI service that calls Etsy search across six regions, computes a 5-signal opportunity score (demand, competition, price floor, seasonality, novelty), and asks a self-hosted Ollama model for an optimised listing strategy and design brief per opportunity.
architecture
external → compute → store → ui
outcome
Keyword-to-listing strategy in minutes, with a regional comparison view that surfaces the best market for the same idea.
stack
capabilities
- →Six-region keyword scoring
- →Five-signal opportunity score
- →AI-generated titles, tags, and pricing strategy
- →Design brief generation
- →Listing optimisation
lessons learned
- 01Self-hosting Ollama instead of paying per-token meant I could run dozens of opportunities in a single session without watching a budget.
- 02Five signals beat ten — adding more inputs made the score noisier without making it more accurate.
- 03The AI is best at reframing, worst at inventing. It expands a keyword into a brief; it should never invent the demand signal.