Patent PreCheck
Patentability pre-check for code — USPTO pillar scores, filing-readiness, and prior-art signals.
README
@patentprecheck/mcp — Patent PreCheck CLI + MCP server
Run a patentability pre-check on your code from inside your terminal or your AI coding agent (Cursor, Claude Code, Codex, Antigravity). It wraps the hosted Patent PreCheck engine, so no API keys are required — the scoring engine and prior-art corpus stay server-side.
Informational only — not legal advice and not a substitute for a licensed patent attorney.
What it does
- Scores code or an invention description across the four USPTO statutory pillars (§101 eligibility, §102 novelty, §103 non-obviousness, §101 utility) plus a separate §112 filing-readiness signal.
- Reports the band (Not Ready → File Ready), the pillar holding the band back, top opportunities to strengthen, and how much prior art was consulted.
- Exposes MCP tools so an agent can score the code it just wrote, inline.
Install
# Run directly (no install)
npx -y @patentprecheck/mcp score ./src/widget.ts
# Or install globally
npm i -g @patentprecheck/mcp
precheck score ./src/widget.ts
CLI
precheck score ./path/to/file.ts # human-readable report
precheck score ./file.ts --format json # raw JSON
cat invention.md | precheck score - # read from stdin
precheck score ./file.ts --min-score 60 # exit 4 if below threshold (CI gate)
precheck pillars # scoring reference (no network)
precheck review # Interactive Code Review signup URL
precheck mcp # run as an MCP server over stdio
Exit codes: 0 ok · 1 usage · 2 request error · 3 §101 gate not passed ·
4 below --min-score. stdout is clean data; progress/errors go to stderr.
Use in an AI coding agent
See config-examples/. Quickest paths:
# Claude Code
claude mcp add patent-precheck -- npx -y @patentprecheck/mcp mcp
// Cursor — .cursor/mcp.json
{ "mcpServers": { "patent-precheck": { "command": "npx", "args": ["-y", "@patentprecheck/mcp", "mcp"] } } }
MCP tools: precheck_score, precheck_pillars, precheck_start_review.
Cursor users can skip the JSON and use the Add to Cursor button at the top.
The server is published to the official MCP Registry
(io.github.Patent-PreCheck/patent-precheck), so MCP-aware clients can discover it directly.
Hosted endpoint (no install)
There's also a hosted, keyless Streamable HTTP MCP server, so remote-capable
clients can connect without npx or a local Node install:
https://patentprecheck.com/mcp
// Cursor — remote server, no command needed
{ "mcpServers": { "patent-precheck": { "url": "https://patentprecheck.com/mcp" } } }
Same three tools and the same scoring engine. The hosted variant takes invention
text inline via code (it never reads files from your machine — use the local
npx server above if you want the path argument).
Develop from source
git clone https://github.com/Patent-PreCheck/patent-precheck-mcp.git
cd patent-precheck-mcp
npm install
node bin/precheck.js pillars
echo "a novel rate limiter that ..." | node bin/precheck.js score - --format text
node bin/precheck.js mcp # stdio server; blocks waiting for an MCP client
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。