How to Find Similar Engineering Solutions in Patents with AI
Find similar engineering solutions already sitting in patents with AI — Patsnap TRIZ Case Library retrieves real, TRIZ-annotated cases from a 45M+ structured patent library by the technical contradiction or effect you describe.
In one example run, a query for “improve heat dissipation without a larger fan or enclosure” returned 10 structured patent cases. Each case includes a problem summary, an effect summary, and TRIZ annotations — but no numeric relevance score and no public patent link, so you read the summaries rather than trust a ranking number.
Example run, August 2026 — one observed instance, not a performance benchmark.When a mechanical or thermal problem in your system resembles something that’s likely been solved before — just not in your industry — the fastest way to find that prior solution isn’t a patent keyword search, it’s a search by the underlying contradiction or physical effect. Claude can help you frame that contradiction, but it doesn’t hold a structured library of patent cases or TRIZ annotations to search against. Patsnap TRIZ Case Library is a retrieval tool — not a concept generator — that matches your stated contradiction, functional relationship, or scientific effect against real patent cases and returns their problem, effect, and innovation summaries alongside TRIZ annotations.
Which questions are a good fit
- “Has anyone solved improving X without worsening Y, in any industry?”
- You can name the two parameters in tension (temperature vs. size, strength vs. weight)
- You want real patent cases to cite as precedent, not new AI-generated concepts
- You want to search outside your own field on purpose (cross-domain discovery)
- You want AI to synthesize a new concept — that’s Solution Engine, a separate tool
- You need a public, citable patent number for every result — this returns internal case IDs
- You need a numeric similarity or relevance score — none is returned
- You need a legal novelty, FTO, or infringement search
Example: compact battery-pack heat dissipation
The problem: improve heat dissipation in a compact battery pack without adding a larger fan or a substantially larger enclosure — a classic TRIZ contradiction between improving temperature control and worsening the volume of a stationary object.
Find cross-industry engineering solutions for improving heat dissipation in a compact battery pack without a larger fan or enclosure.Describe the contradiction like this — improving what, without worsening what — and Claude maps it to the tool’s standard parameter list and runs the search. In the recorded run, the query returned 10 cases; the top three carried case IDs be012cf4-94bc-44d2-af98-d4d968aa6571, be740981-3ce2-41e9-9256-47fc9225ccb5, and ef9e2e63-9b34-4b61-bee9-d5429f643af7.
| Rank | Case ID | Fields returned per case |
|---|---|---|
| 1 | be012cf4-94bc-44d2-af98-d4d968aa6571 | Problem summary, effect summary, innovation summary, and TRIZ annotations (technical contradiction, SVOP, scientific effects) |
| 2 | be740981-3ce2-41e9-9256-47fc9225ccb5 | |
| 3 | ef9e2e63-9b34-4b61-bee9-d5429f643af7 |
These are Patsnap’s internal case identifiers, not public patent numbers — there’s no external link to click through and verify independently.
This tool returns cases sorted by relevance, but doesn’t expose a numeric score — you judge fit by reading each case’s problem/effect summary, not by a ranking number. Any industry you attribute to a case (e.g. “this came from refrigeration”) is inferred from the summary text, not a labeled database field — describe it as an inference, not a confirmed fact, when you write it up.
How to run it
https://connect.patsnap.com/37b2d2/mcp?apikey=yourapikey Replace yourapikey with your own Patsnap Open Platform API key before connecting. Don’t paste the completed link into a public chat, a screenshot, or a code repository — anyone with it can use your credits.
- Get a Patsnap Open Platform API key.
- Copy the connection URL or JSON config from the Connect panel on the TRIZ Case Library page.
- Add the server to Claude Desktop or another MCP-compatible client, then restart it.
Once connected, describe your contradiction to Claude in plain language, as shown above — no need to know the tool’s internal parameter IDs yourself.
How to read and use the results
Read the problem and effect summary for each case, not just its rank position, and note which TRIZ annotation (technical contradiction, SVOP, or scientific effect) matched your query — that tells you why the case surfaced. Ask yourself whether the transfer makes physical sense in your context: a solution from an unrelated industry can be genuinely useful, but only if the underlying mechanism — not just the surface description — actually applies to your materials, scale, and operating conditions.
- Retrieving cases matching your contradiction, effect, or keyword
- Returning problem/effect/innovation summaries per case
- Attaching TRIZ annotations to each result
- Translating your problem into search terms
- Summarizing and comparing returned cases
- Drafting a transfer hypothesis to test
- Verifying the transfer hypothesis experimentally
- Checking the case’s original assumptions hold in your context
- Any patent-clearance review if a result sits close to your design
Search parameters and tool details (optional)
The server exposes two tools. search_triz_case requires at least one of five criteria — technical_contradiction (an improving/worsening parameter pair from a standard list of 39, optionally with an invention principle), svop (subject/verb/object/parameter function description), efficacy, oxford_effects (scientific effects), or keyword (up to five terms) — and an optional filters object for authority, legal status, applicant, and IPC/CPC classification, including an exclude option specifically for cross-domain discovery. list_triz_search_terms looks up the valid spelling for efficacy or scientific-effect terms before you search with them.
Each returned case includes its ID, problem/effect/innovation summaries, and TRIZ annotation data. There’s no separate “fetch full case” tool — the search call itself returns the summary content, so no follow-up call is needed to read a case in more depth.
Find cross-industry engineering precedent
Connect Patsnap TRIZ Case Library and search real patent cases by contradiction or effect.
Frequently asked questions
How is this different from Patsnap Solution Engine?
TRIZ Case Library retrieves existing patent cases matching your contradiction — it doesn’t generate new synthesized concepts. Solution Engine takes a problem statement and produces new candidate concepts using TRIZ reasoning and case matching together.
Why isn’t there a relevance score or a public link for each case?
The tool sorts by relevance internally but doesn’t expose a numeric score, and case IDs are Patsnap’s internal identifiers rather than public patent numbers. Judge fit by reading the returned summaries.
Do I need to know the standard TRIZ parameter numbers to use this?
No. Describe your contradiction in plain language and Claude maps it to the tool’s parameter IDs. Keyword or effect-based search is also available if you don’t have a formal contradiction framed yet.
Disclosure & disclaimer
Commercial disclosure and sources
Published by Patsnap, which develops and sells the Patsnap TRIZ Case Library MCP server described here — an editorial tutorial, not an independent review. Server capabilities and connection details come from the product page and Patsnap Open Platform documentation as accessed August 2026; the example is one recorded run, not a repeated benchmark.
Engineering limitations
Retrieved cases are research leads for engineering exploration, not validated solutions, a novelty search, or a legal opinion. Verify any transfer hypothesis experimentally, and confirm patent status independently before relying on a specific case for IP purposes.
Trademarks, product names, and company names mentioned — including Claude and Cursor — belong to their respective owners and are used here for identification only.
