About
Most product research is just a list.
Dropship Spy is an AI Product Testing OS. It exists to answer one question honestly — is this product worth spending money on? — and then to get you from that answer to a real test without a fortnight of work in between.
The problem
Two ways to lose money on a product test.
Spending first, checking later
The usual sequence is a hunch, a product page, a creative, and a few hundred dollars of ad spend used as the research method. By the time the data arrives, it has already been paid for — and most of what it tells you could have been known beforehand.
Databases of yesterday's winners
The tools built to fix that mostly ship a product database — the same curated feed, sold to everyone, recycling products that were already saturated when they were added. Picking from a list that thousands of other stores are picking from isn't an advantage, and it doesn't tell you a thing about your margins.
The idea
An operating system for testing products
A product test has three moving parts: deciding whether the product deserves a test at all, making the creative that carries it, and building the page the traffic lands on. Those three jobs normally live in three different tools, and the handover between them is where days disappear.
Dropship Spy puts them on one rail and one currency. You start from a product you chose — not one we pushed at you — and every step spends from the same credit balance. Research is 1 credit, an AI video ad is from 1 credit, a landing page is 2. Nothing publishes itself and nothing spends ad budget on your behalf: the output is a decision and the assets to act on it.
The research is deliberately built on current signals rather than a stored catalogue. Each run maps the data points that matter for that specific product, which is why two searches don't look identical — and why the answer changes when the market does. What the analysis covers goes through that in detail.
The workflow
Three steps, one balance.
01
Research
1 creditPaste a product. Dropship Spy analyses a changing mix of current demand, sourcing, pricing, competition, buyer-intent and risk data points, then returns a verdict — Worth Testing, Caution or Skip — with a private evidence trail behind it.
02
Ad Studio
from 1 creditPick a style — UGC, cinematic, unboxing, problem-solution — and an engine from Wan 2.5 and Hailuo through Kling and Seedance up to Google Veo 3.1 with native audio. AI writes the shot script and renders a vertical-ready MP4. Upload a real product photo and the ad is built around it.
03
Funnel Builder
2 creditsAI designs a conversion-focused landing page around the product with the ad embedded, and hands you one self-contained HTML file. No Dropship Spy branding, no lock-in — host it on Shopify, Vercel, Netlify or any static host.
Each step is broken down further across the feature pages. Agents can drive the research step too — Dropship Spy runs a hosted MCP endpoint, so Claude, Claude Code or Cursor can spend your credits on research with a scoped OAuth connection, and nothing else.
How we work
Evidence first, verdict second.
If you want the mechanics rather than the principle, the numbers behind the verdict lays out the receipts, the coverage and the margin arithmetic behind one decision.
Every run leaves receipts
A completed research includes a private evidence trail showing the coverage behind its verdict. You can see what the call was made on rather than taking a score on faith.
The answer is allowed to be no
Skip is a real verdict, not a formality. A tool that only ever finds winners isn't researching anything — it's selling you permission to spend.
The maths has to work
The margin reality check puts source cost, shipping and 15–20% ad spend against your retail price. Plenty of products that look good on demand fall over here.
You can argue with it
Every research is a thread. Ask for a faster supplier, a different angle, a UK comparison. Five follow-ups come with each run, and each one after that costs a credit — a follow-up re-runs the full engine rather than replaying a cached answer.
Why there's no free tier
Paid from day one, on purpose
There is no free trial and no free credit. Every plan is paid from day one, starting at $29/month for 25 credits.
We took the free credit away on purpose: it paid for people to run one search on a demo product and disappear, and the cost of that came out of making the product better for the people who actually launch. Real research costs real money to run — every credit calls live data providers and video engines — so the choice was between subsidising drive-by searches and spending that money on the people using the thing.
What that buys you: the same engine on every credit, unused credits that roll over, automatic refunds on failed renders, and a plan you can cancel at any time with the credits you've paid for staying yours. If you want to try the thinking before you pay, the free tools are genuinely free and need no signup.
What we don't do
- No database of millions of products to browse — a list everyone else is also browsing is not an edge.
- No guaranteed winners. The verdict is generated by an LLM against live third-party data and it can be wrong; the decision stays yours.
- No lock-in on what you build. Funnels download as plain HTML and the MP4s are yours.
- No testimonials, star ratings or customer counts on this site that we can't stand behind.
The full position on outputs, credits and refunds is in the terms, and what we store is in the privacy policy.
Test the next product before you fund it.
Plans start at $29/month for 25 credits — research, AI video ads and landing pages from one balance. Questions first? The FAQs cover most of them, and our inbox covers the rest.