Features
Dropship Spy is an AI Product Testing OS. Three steps, one credit balance: check whether a product is worth testing, make the ad, build the page. Nothing publishes and nothing spends on your behalf.
1 credit
Paste a product name or URL. Dropship Spy analyses a product-specific mix of current demand, sourcing, pricing, competition, buyer-intent and risk data points, then returns one verdict — Worth Testing, Caution or Skip — with a private evidence trail behind it.
How the analysis worksfrom 1 credit
Pick an ad style and a video engine. AI writes the shot script, the engine renders it, and you download a vertical-ready MP4 for TikTok, Reels or Shorts. Upload a real product photo and the ad is built around that photo, not a lookalike.
Engines and ad styles2 credits
AI writes and designs a conversion-focused landing page around the product, with your generated ad embedded in the hero. The output is one self-contained HTML file — no branding, no lock-in.
Where the file goesWhat each part does, what it will not do, and what it costs.
The data points behind a run, what each verdict means, and why an unverifiable number comes back as “couldn’t verify” instead of a guess.
Read moreSource receipts, coverage and confidence, and the margin reality check that prices in source cost, shipping and a 15–20% ad-spend assumption.
Read moreThe shot script, on-screen hook and social caption written per ad style — plus the headline, offer and FAQ copy the Funnel Builder produces.
Read moreSix render engines from Wan 2.5 to Google Veo 3.1 with native audio, six ad styles, vertical output, and an automatic refund when a render fails.
Read moreOne portable HTML funnel you host anywhere, and MCP agent access so Claude or another AI agent can run research over a hosted endpoint with OAuth.
Read moreEach completed run records which sources were checked, when, how many results they returned, and the limitation on each one. It is private to your account.
Source price plus shipping plus a 15–20% ad-spend assumption — stated as an assumption, not as observed performance.
5 follow-up questions are included with every research, then 1 credit each. Ask for a different supplier, another ad angle, or a margin stress-test.
If a video job comes back terminally failed, its credits are returned to your balance automatically — once, never twice.
MP4 ads and one-file HTML landing pages download to your machine. Nothing is locked to a hosted builder.
Connect Claude or another MCP client to the research tools with a browser OAuth approval. No API keys to copy around.
Plans differ only in how many credits they include — see the three plans. Agent access is documented on the agents page.
No. Each step is a separate action with its own credit cost: research is 1 credit, an AI video ad is 1–8 credits depending on the engine and length, and a landing page is 2 credits. Run research on its own, or go straight to Ad Studio for a product you have already decided on.
One of three calls — Worth Testing, Caution or Skip — with a one-line reason, a suggested retail price, an estimated source price, an estimated margin after ads, a demand trajectory and a saturation read. Anything the sources could not confirm comes back as “couldn’t verify” rather than a guess, and thin evidence pushes the verdict to Caution rather than a confident call.
Yes. The Funnel Builder outputs a single self-contained HTML file — all styling inline, no external frameworks, no Dropship Spy branding. Host it on Shopify, Vercel, Netlify or any static host.
The credits are refunded automatically. When a job comes back terminally failed, the failure and the refund are written in one transaction, so overlapping page refreshes cannot double-refund or miss it.
Plans start at $29 a month for 25 credits. There is no free trial and no free credit — every plan is paid from day one, and unused credits roll over.
Prefer to read first? The blog covers product testing method in more depth.
