How to Check Design Patent Infringement Risk with AI
Before you launch a product, screen its design against existing patents with AI — upload a photo to Design Patent Infringement Search and get back a ranked list of visually similar designs to review.
This workflow works in Claude, Claude Code, ChatGPT, or Codex when the selected client supports the required MCP connection. After connecting, the research steps and evidence boundaries are the same.
Cross-border e-commerce teams launching a new product often need to check whether its design could conflict with existing design patents in target markets — and doing that properly means comparing an actual product photo against a global body of protected designs, not running a text search. An AI assistant or coding agent can walk through the screening logic with you, but on its own it has nothing to compare your image against. Design Patent Infringement Search closes that gap — it’s a Patsnap Open Platform server built specifically for image-led design screening, part of the same platform that connects AI agents to Patsnap’s broader patent, R&D, and life-sciences data.
Why this needs Design Patent Infringement Search
An AI assistant or coding agent can suggest which jurisdictions to check and which status filters to apply, but without a design-patent database and image-similarity engine it can’t retrieve or rank actual matching designs. This server compares a product image against a database the vendor describes as covering 190M+ patents across 171 countries and regions, returning a ranked list of visually similar designs with jurisdiction, status, similarity scores, and tiered risk signals. Its role is image-led screening before human IP review — not a final infringement opinion.
It works through three tools: submit_workflow accepts an image URL plus optional country, status, date-range, and classification filters and starts an asynchronous screening task; get_task_status lets Claude poll for completion; get_stage_result retrieves the ranked comparisons once ready.
Prepare the product image and screening scope
The required input is a product image URL. A publicly accessible link is the most reliable option; if your image is a local file, upload it somewhere accessible first. Whether a base64/data URI works directly depends on your specific MCP client’s handling of the parameter, not a documented limitation of the tool itself — check your client if you want to skip external hosting.
Decide which jurisdictions to screen using the tool’s own country codes — for example ["CN", "US", "EP"]. Note that a code here reflects the tool’s filter, not necessarily a single unified design-right system: design protection frameworks differ by authority (national offices, regional systems, and international registrations all work differently), so confirm what a given code actually covers for your target market before treating a result as market-specific clearance. The simple_legal_status filter uses numeric codes — 0 for expired, 1 for active, 2 for pending, comma-separated, defaulting to “1,2” if omitted. An optional date range (apply_start_time/apply_end_time, format YYYYMMDD) narrows by filing date, and loc_items accepts structured Locarno classification entries if you have them — separate from main_field, which is just a free-text product description, not a classification code.
For this tutorial, the input is a public smartwatch image, jurisdictions US and EP, legal status active and pending (the default), and no classification or date filter applied.
Connect Design Patent Infringement Search
https://connect.patsnap.com/937ec6/mcp?apikey=yourapikey Replace yourapikey with your own Patsnap Open Platform API key before connecting. Keep the completed URL private. - Claude: add the remote MCP connection from Claude’s connector or integration settings.
- Claude Code: add the server as a remote HTTP MCP connection in your user or project configuration.
- ChatGPT: add the server through the supported connector or developer-mode MCP setup available to your workspace.
- Codex: add the server in Codex’s MCP configuration for the relevant local workspace.
The interface differs by client, but you use the same Patsnap connection URL and the same natural-language research prompt.
- Get a Patsnap Open Platform API key.
- Copy the connection URL or JSON config from the Connect panel on the Design Patent Infringement Search page.
- Add the server to your AI assistant or coding agent, Cursor, or another MCP-compatible client, following that client’s own setup steps.
- Confirm the workflow, status, and result tools are available before you ask.
Submit and check the screening task
Screen this smartwatch image for design-patent risk in the US and EP, active and pending patents only: [HTTPS image URL]Claude calls submit_workflow with the image URL and your filters; the server returns a task ID and begins processing asynchronously. Because completion time depends on jurisdiction count, patent-set size, and server load, Claude polls with get_task_status rather than waiting inline — ask it to check progress, and once the similarity-search and feature-comparison stages report complete, it retrieves the comparison results with get_stage_result.
Read the ranked results
| Field | Value |
|---|---|
| Comparisons returned | 16 |
| Top result | USD933065S |
| Similarity score | 0.96 |
| Risk signal | High |
| Processing time | 142.68 seconds, zero processing failures |
A feature-comparison view highlights which visual elements drove the similarity score. Higher similarity plus an active or pending status in a target jurisdiction raises the tiered risk signal, which prioritizes results for human review — it doesn’t determine infringement.
Continue to human IP review
- Rerun with additional jurisdictions such as CN or JP if launching in those markets.
- Narrow the date range to recent filings if you want to prioritize newer grants.
- Ask your AI assistant or coding agent to retrieve the feature-comparison data for the top few results so a reviewer can see which visual elements drove the ranking.
Each follow-up uses the same connection, letting the team iterate and hand off a complete evidence package to legal counsel for final review.
Screen a product image for design risk
Connect Design Patent Infringement Search and rank visual matches before launch.
Frequently asked questions
Why can’t I just Ask your AI assistant or coding agent to search design patents directly?
Claude has no live connection to design-patent records or an image-similarity engine. Design Patent Infringement Search supplies that connection, retrieving and ranking matches based on your product image.
What image input does the tool require?
A product image URL. A publicly accessible link works most reliably; whether other formats work depends on your MCP client, not a fixed limitation of the tool.
How do I interpret the similarity score and risk signal?
The similarity score measures visual overlap with a design-patent drawing; the tiered risk signal combines similarity, status, and jurisdiction to prioritize review. Neither is a legal determination of infringement or clearance.
What should I do after getting ranked results?
Move high-priority results to legal review: compare the product’s overall visual impression to the claimed design, verify current status in the official register, and consult IP counsel before a launch or design-change decision.
Can I use this workflow in Claude, Claude Code, ChatGPT, or Codex?
Yes, provided the selected client and workspace support the required MCP connection. Connect the same Patsnap server, then use the natural-language steps shown in this guide; interface details may differ by client.
Disclosure & disclaimer
Who published this
Published by Patsnap, which develops and sells the Design Patent Infringement Search MCP server described here. This is an editorial tutorial, not an independent review.
How the information was gathered
Server capabilities and connection details are taken from the Design Patent Infringement Search product page and Patsnap Open Platform documentation, as accessed on August 18, 2026. They are not the result of independent testing beyond the recorded screening example in this tutorial. Coverage figures and product capabilities change frequently and may have changed since publication.
Not legal advice
Similarity and risk signals in this tutorial are a screening aid, not a legal opinion on infringement, validity, or freedom to operate. Consult qualified IP counsel before making a launch, sourcing, or design-modification decision.
Trademarks
All trademarks, service marks, product names, and company names mentioned — including Claude and Cursor — are the property of their respective owners and are used here solely for identification purposes. Their use does not imply any affiliation with, sponsorship by, or endorsement from their respective owners.
