ClaimHit
Find products & standards using a patent - verified targets for licensing or enforcement.
Installation
Ajoutez ceci à la configuration MCP de votre client (Claude Desktop : claude_desktop_config.json ; Cursor : .cursor/mcp.json ; VS Code et les autres utilisent la même forme « mcpServers »).
claude mcp add --transport http claimhit 'https://www.claimhit.com/api/mcp'{
"mcpServers": {
"claimhit": {
"type": "http",
"url": "https://www.claimhit.com/api/mcp"
}
}
}Documentation
Tirée du README du projet sur GitHub, actualisée chaque semaine. Les liens ouvrent GitHub.
Sommaire
ClaimHit — MCP Server
Find the companies using a patent — to license, partner, or enforce — right from your AI client.
ClaimHit is a patent licensing and infringement research engine. Give it a patent number and it returns real, currently-shipping products (and, for SEPs, standards) that use the claimed technology — each one verified against live evidence on the manufacturer's own pages, not guessed. It also generates element-by-element claim charts for the targets you want to pursue.
This is the remote MCP server for claimhit.com, so you can run a search, work through the results and build charts without leaving Claude (or any MCP client).
Connect
- URL:
https://app.claimhit.com/api/mcp - Auth: OAuth 2.1 — you sign in to your ClaimHit account and approve; the connector then works with your account's credits. No key to paste, no config file.
In Claude: Settings → Connectors → Add custom connector → paste the URL → Connect → sign in → Authorize.
You'll need a (free) ClaimHit account. A free account includes a trial search so you can try it immediately; deeper usage is credit-metered on your plan.
API keys
For a programmatic client that cannot do OAuth, generate a key (prefix ch_) at
app.claimhit.com/claimhit/account → Developer
tab, and send it as a Bearer token:
Authorization: Bearer ch_live_…A ?api_key= query parameter is also accepted, for clients built against it. Prefer the
header wherever you have the choice: a credential in a URL is a credential in access logs
and in browser history.
A ch_test_ key is a sandbox key — it returns fixture data and spends no credits.
Tools
| Tool | What it does | Cost |
|---|---|---|
claimhit_search | Search a patent number for infringing products / essential standards. Returns verified, evidence-backed targets. | 1 search credit |
claimhit_get_result | Read a past search — and narrow, filter, sort and page through it. Ask for one company, a technology, the low-confidence leads, a different order, the next page, or the full evidence and per-element scores behind a score. | Free, unlimited |
claimhit_refine_product | Look again at one target. Give it a vague result plus what you want instead — a specific model, the latest one, or a capability or market — and it returns up to five real, current products with a reason for each, grounded in a live web search. | Free |
claimhit_generate_chart | Generate an element-by-element claim chart mapping a patent claim to a product or standard. Quotes the cost and the target it matched first; runs nothing until confirmed. | 1 chart credit, confirmed |
claimhit_rerun_search | Re-run a prior search to pick up newly-released products, repeating the inputs that steered it. Lists what it will repeat before running; can widen a scoped search to the whole market. | Free within 6 months, confirmed |
claimhit_get_history | List your recent searches. | Free |
claimhit_get_credits | Check your remaining search and chart credits. | Free |
One search is the paid step; the conversation after it is not. Narrowing, sorting, paging, reading evidence and refining a target are free and uncapped, and the two tools that spend a credit quote the cost first and run nothing until you agree — so a long session working through a result cannot run up a bill because the assistant decided to be thorough.
A search also completes in phases: claimhit_get_result returns a status, and background
enrichment keeps finding implementers for several minutes after the first results land.
Treat results as final only at COMPLETE.
How it works
ClaimHit doesn't just ask an LLM "who infringes this?" — several models run the discovery pass independently, and every candidate is then put through claim-alignment scoring and an evidence check (the product must resolve on the manufacturer's own domain and match the claim vocabulary), plus a forward-citation expansion that surfaces companies already active in the technology. The goal is an honest, comprehensive target list you can take into a licensing campaign — not a list of plausible-sounding guesses.
Results are scored across four independent factors:
- Model consensus — how many models independently flagged the target
- Claim coverage — which claim elements are matched, weighted toward inventive steps
- Evidence strength — specific datasheets and regulatory filings score higher than inferences
- Functional equivalence — a doctrine-of-equivalents proxy
Full methodology at claimhit.com/methodology.
SEP analysis
Standards mode searches published specifications from 3GPP, IEEE, ETSI, ITU-T and IETF, citing the clause where the relevant requirement appears and marking each requirement mandatory, optional or conditional.
What this is not
ClaimHit returns a probable list of products with the evidence behind each one. It is preliminary research to focus attorney time — not a legal opinion, and not legal advice.
Troubleshooting
Connection fails with a redirect (307 / 308)
Use https://app.claimhit.com/api/mcp. The marketing hosts claimhit.com and
www.claimhit.com redirect to it, and MCP clients that do not follow redirects cannot
connect through them.
Authentication failed
Re-authorize the connector, or — if you are using a key — check that it starts with ch_
and is still active. Generate a new one at
app.claimhit.com/claimhit/account → Developer
tab.
No results returned
Check your balance with claimhit_get_credits, and top up at
app.claimhit.com/claimhit/credits. If a search
did run, try claimhit_get_result with risk: ["LOW"] — the low-confidence leads are
excluded from the default list.
Fewer results than expected
The default list is HIGH and MEDIUM confidence only. Pass risk: ["HIGH","MEDIUM","LOW"]
to see everything, and page with offset — the response's more field says when further
matches remain.
Patent not found
Include the kind code: US10123456B2, not US10123456.
Support Email hello@claimhit.com with your request ID.
About
Built by Scintillation Research and Analytics Services Pvt. Ltd.
- Product: https://www.claimhit.com
- Security: https://www.claimhit.com/security
- Privacy: https://www.claimhit.com/privacy
Questions fréquentes
Qu'est-ce que ClaimHit ?
Find products & standards using a patent - verified targets for licensing or enforcement.
Comment installer ClaimHit ?
Il fonctionne comme serveur distant. Dans Claude Code : claude mcp add --transport http claimhit 'https://www.claimhit.com/api/mcp'. Pour les autres clients, collez la configuration JSON de cette page dans leurs réglages MCP.
Faut-il une clé API pour ClaimHit ?
Son entrée au registre n'indique aucune clé ni secret obligatoire. Certaines fonctions peuvent tout de même demander un compte auprès du service auquel il se connecte.
ClaimHit est-il open source ?
Son code est public sur GitHub, mais le dépôt ne déclare pas de licence standard.
Qui maintient ClaimHit ?
Il est publié par Bikram2112. AI.tn le référence sans le maintenir ni le recommander.
Issu du registre MCP officiel. Être listé n'est pas une recommandation : vérifiez le code et les permissions d'un serveur avant de lui donner accès à vos données.