How to Find Expired Patents, Manufacture the Products, and Sell Them on Amazon
I created a course/community to master Openclaw and Claude AI agent automations for beginners to start making money and selling to companies
Check it out here: https://www.skool.com/raycfu/about
WHY THIS WORKS
Most people think of patents as legal shields. Something lawyers deal with. But a patent is also an engineering document. To get a patent granted, you have to disclose enough technical detail that someone skilled in the field could reproduce the invention. Dimensions. Materials. Assembly steps. Performance specs. That is the entire point of the patent system. You get temporary protection. In exchange, you give the public a complete blueprint.
When the patent expires, the blueprint stays. It becomes a free manufacturing manual. But nobody reads them because patent documents are written in dense legal language that looks like gibberish to normal people. Claude reads them and tells you what the product actually does in plain English.
The products you are looking for are not failed inventions. They are products that sold well, had real demand, and solved real problems but simply stopped being made because the company behind them died. The product did not fail. The business did. And nobody went back to check.
STEP 1: ACCESS THE PATENT DATABASE
The USPTO publishes everything through their Bulk Data portal. Full text, full drawings, full claims. Every detail you need to rebuild a product from scratch. It is free and public.
USPTO Bulk Data: https://bulkdata.uspto.gov
You can also search individual patents on Google Patents which has a cleaner interface and lets you filter by expiration status and see the full filing with drawings.
Google Patents: https://patents.google.com
For your first search, go to Google Patents and use these filters:
Status: Expired Type: Utility (skip design patents, they only cover appearance) Date range: Expired after 2014 (recent enough that the market still exists) Categories to focus on: household, tools, pet products, office supplies, garden, kitchen
You specifically want patents from small to mid-size companies. Not IBM. Not Samsung. Not Qualcomm. Small companies where the tech is simple enough to manufacture independently. Consumer products. Hardware. Tools. Home goods.
STEP 2: BUILD YOUR SCRAPING AND FILTERING PIPELINE
Manually browsing patents one by one does not scale. You need to pull them in bulk and filter programmatically.
Write a simple Python scraper that pulls expired utility patents from the USPTO Bulk Data API. If you do not code, you can hire someone on Upwork for $50 to $100 to build this for you. The filter logic is straightforward.
Filter for: Status: expired Type: utility only Assignee size: small companies (no mega-corps) Categories: household, tools, pet products, office supplies, garden, kitchen Expired after: 2014 Minimum claims: 3 (too few claims means the patent is too narrow) Maximum claims: 25 (too many claims means the product is too complex to manufacture cheaply)
This first filter will give you roughly 300,000 to 400,000 candidates. Still way too many to read manually. That is what Step 3 is for.
To convert the raw patent files into something Claude can actually process, use markitdown. It strips any document format down to clean Markdown.
markitdown: https://github.com/microsoft/markitdown
For batching the files into prompt-ready chunks, use files-to-prompt.
files-to-prompt: https://github.com/simonw/files-to-prompt
STEP 3: SCORE PATENTS WITH CLAUDE
This is where the pipeline becomes powerful. You feed patents to Claude in batches of 50 and it scores each one for commercial viability.
Use this system prompt:
ROLE: Patent Commercial Viability Analyst
INPUT: Expired US utility patent (full text + claims)
ANALYZE AND RETURN:
- PLAIN_ENGLISH: What does this actually do?
- CONSUMER_VIABLE: Could a consumer version exist? (yes/no)
- BOM_ESTIMATE: Bill of materials at 1000 MOQ (Alibaba pricing)
- AMAZON_GAP: Does any current listing use this exact mechanism?
- REVIEW_SIGNAL: What do competing product reviews complain about?
- SCORE: Commercial viability 1-10
REJECT IMMEDIATELY: Requires FDA/FCC clearance Needs custom semiconductor fab Chemical formulation patents Software/algorithm patents Requires tooling over $50K
RETURN FORMAT: JSON only. No commentary.
Each batch of 50 patents takes about 90 seconds to process. The results are mostly noise. Score 1 after score 1. Old semiconductor layouts. Obscure chemical formulations. Telecom protocols from 2004. Medical imaging tools that need FDA clearance.
But roughly 1 in 80 comes back with a score of 7 or higher. Those are the ones worth investigating.
A high-scoring result looks like this:
Patent ID: US8,234,811 Plain English: Self-watering planter insert with passive felt wick. Pulls moisture from reservoir to soil via capillary action. No pump, no battery. Consumer viable: yes BOM estimate: $1.60-2.10 at 1000 MOQ Amazon gap: yes Review signal: Competing products have water pooling at bottom, wicks clog after 2 weeks, not optimized for small pots Score: 8
STEP 4: VALIDATE THE WINNERS
For every patent that scores 7 or higher, do manual validation before spending any money.
Check Amazon for the exact mechanism. Go to Amazon and search the product category. Not "is there a similar product" but "does any current listing use this exact patented mechanism?" There will probably be similar products on the market. That is good because it proves demand exists. What you are looking for is whether anyone is using this specific design. If nobody is, you have a gap.
Read the reviews on competing products. This is the most valuable 30 minutes you will spend. The reviews on the top 5 competing products tell you exactly what is wrong with everything currently on the market. If the patented design solves those exact complaints, you have a winner.
For example, competing self-watering planters have reviews saying "water pools at the bottom," "my basil still died," and "wick stopped working after 2 weeks." The patented design solved exactly those problems because the inventor clearly tested it. And the patent document says so with diagrams, measurements, and material specifications.
Verify the patent is actually expired. Go to Google Patents, search the patent number, and confirm the status shows expired. Check the expiration date. Check that no continuation patents were filed that might still be active. This takes 5 minutes and is not optional.
Estimate the market size. Search the product category on Amazon and look at monthly search volume using a tool like Jungle Scout or Helium 10. You want categories with at least 50,000 monthly searches. Below that and the market is too small to justify a production run.
STEP 5: GET MANUFACTURING QUOTES
This is the part most people do not realize is possible. A patent document is literally a manufacturing instruction manual written in legal language. Dimensions. Tolerances. Materials. Assembly order. Everything a factory needs to quote production.
Go to Alibaba (alibaba.com) and search for manufacturers in the relevant product category. Contact 3 to 5 manufacturers. Send them the patent drawings directly. You do not need to explain the product in your own words. The patent drawings do that for you.
Ask for: Unit price at 500 MOQ (minimum order quantity) Unit price at 1,000 MOQ Unit price at 5,000 MOQ Mold or tooling costs if applicable Sample cost and timeline Production timeline for first order Material options and any cost differences between them
Expect your first quotes within 3 to 5 business days. Compare across all manufacturers. The cheapest quote is not always the best. Look at communication quality, sample willingness, and whether they ask smart questions about the design (that means they actually read it).
Order samples before committing to a production run. Always. Samples cost $20 to $100 typically plus shipping. Test the product yourself. Compare it to what is currently on Amazon. Make sure the patented design actually delivers the improvement the patent claims it does.
STEP 6: CALCULATE YOUR UNIT ECONOMICS
Before you order a production run, know your numbers cold. Here is the math for a typical product:
Manufacturing cost per unit (at 1,000 MOQ): varies by product Packaging and labeling: $0.30 to $0.80 per unit Shipping to Amazon FBA warehouse: $0.50 to $1.50 per unit depending on weight and origin Amazon FBA fulfillment fee: $2.50 to $5.00 per unit depending on size Amazon referral fee: 15% of sale price PPC advertising (estimated): $1.00 to $2.00 per unit in the first 3 months Returns and damages: budget 3 to 5% of units
Add all of that up. Subtract from your selling price. That is your net margin per unit.
The math from one of the real products found through this system:
Cable clip 30-pack example: Production (30 clips at $0.12 each): $3.60 Packaging and label: $0.40 Shipping to FBA: $0.85 Amazon FBA fee: $3.20 PPC estimated: $1.10 Total cost: $9.15 Sale price: $11.99 Net margin: $2.84 (23.7%) At 800 units per month: $2,272 per month net
Self-watering planter insert: Manufacturing: $1.80 Target sale price: $11.99 Projected margin after PPC and FBA fees: 44%
You want a minimum of 20% net margin after all costs. Below that and one bad month of PPC or returns wipes out your profit. Above 30% and you have a real business.
STEP 7: LAUNCH ON AMAZON
Create your Amazon Seller Central account if you do not have one. The Professional plan is $39.99 per month and is required if you plan to sell more than 40 units per month.
Product listing: Write a title that includes the primary search keywords. Use all 5 bullet points to describe what the product does and why it is better than alternatives. Include the key features from the patent that solve the problems you found in competitor reviews. Do not mention the patent. Nobody cares that it is patented. They care that it works.
Photos: Get professional product photos. This is not optional. Hire a photographer on Fiverr for $50 to $150 or use a product photography service. You need a main image on white background, lifestyle images showing the product in use, and close-up shots of the key features that differentiate you from competitors.
Ship your first batch to FBA: Send your initial production run to an Amazon FBA warehouse. Amazon handles storage, picking, packing, shipping, and returns.
Launch PPC: Start with automatic targeting campaigns at $15 to $25 per day. Let it run for 2 weeks. Then analyze which keywords are converting and move those to manual campaigns. Cut keywords that are spending without converting.
Get initial reviews: Use Amazon's Vine program (available through Seller Central) to get your first 15 to 30 reviews. Reviews are the single most important factor in whether your listing survives the first 90 days.
STEP 8: SCALE AND REPEAT
Once your first product is profitable and running, repeat the entire pipeline. You already have the scraper, the filter, the Claude scoring prompt, and the manufacturer relationships. Finding the next product takes days, not weeks, because the infrastructure is already built.
The real value compounds over time. As you run more patents through the pipeline, Claude's scoring gets more useful because you can refine the prompt based on which products actually worked versus which ones scored high but flopped. Your Alibaba relationships get stronger so you get better pricing and faster turnaround. Your Amazon account builds history so new listings rank faster.
Target 3 to 5 products in year one. Not all of them will work. Expect 1 or 2 out of every 5 to be profitable enough to keep running. That is normal for physical products on Amazon. The ones that work make up for the ones that do not.
TOOLS AND COSTS SUMMARY
USPTO Bulk Data (patent database): Free Google Patents (search and verify): Free markitdown (document conversion): Free, open source files-to-prompt (batch processing): Free, open source Claude API (scoring pipeline): $20 to $50 per scoring run of a few thousand patents Alibaba (manufacturer sourcing): Free to search and message Product samples: $20 to $100 per product Mold/tooling (if applicable): $500 to $5,000 one-time depending on complexity First production run (1,000 units): $1,000 to $3,000 depending on product Product photography: $50 to $150 Amazon Professional Seller account: $39.99/month Initial PPC budget: $15 to $25/day for the first 90 days
Total startup cost for your first product from patent search to Amazon listing: roughly $2,000 to $5,000 depending on the product complexity and production run size. Most of that is the manufacturing cost which you get back when the units sell.
WHAT TO AVOID
Do not target products that need FDA clearance (anything medical, food-contact, or supplement-related). The regulatory cost alone will kill your margins.
Do not target products that need FCC certification (anything electronic that transmits or receives signals). Same reason.
Do not target products with tooling costs over $50,000. The upfront investment is too high for your first product.
Do not skip the sample stage. Ever. A product that looks great on paper can have manufacturing defects that only show up in the physical version.
Do not assume an expired patent means no legal risk. Check for continuation patents, divisional patents, and related patents that might still be active. Google Patents shows related filings. When in doubt, spend $200 to $300 on a patent attorney consultation for a specific product before committing to a production run.
Do not try to sell the "cheapest" version. The products that win on Amazon are the ones that solve specific complaints from competitor reviews. Price for quality. The margins are better and the reviews are better.
THE FULL PIPELINE AT A GLANCE
USPTO Bulk Data API: Pull expired utility patents filtered by category, assignee size, and date.
Python scraper: Filter down to 300,000 to 400,000 candidates matching your criteria.
markitdown: Convert patent files to clean Markdown.
files-to-prompt: Batch patents into prompt-ready chunks of 50.
Claude scoring pipeline: Score each batch on commercial viability. Filter for scores of 7 or higher. About 1 in 80 passes.
Google Patents: Verify expiration status and pull drawings for winners.
Amazon review analysis: Read competitor reviews to confirm the patented design solves real complaints.
Alibaba: Send patent drawings to 3 to 5 manufacturers. Get quotes. Order samples.
Unit economics: Calculate all-in cost per unit versus selling price. Minimum 20% net margin.
Amazon FBA: List the product, ship inventory to FBA, launch PPC, get initial reviews through Vine.
Scale: Repeat with the next product. Infrastructure is already built.
4 million expired patents. Each one a detailed instruction manual for a product that once worked. Most of them are genuinely useless. But some of them are products that were once valuable enough to spend $15,000 protecting and that nobody remembered to keep making. Those are the ones you are looking for.
You just read the full playbook. Most people will close this tab and never implement it. The ones who do usually hit a wall around the technical setup and quit.
Inside the Skool, I walk you through the exact build step-by-step, troubleshoot your setup live in the community, and share the scripts and templates I use to actually land paying clients.
If you want the shortcut instead of the long way around: