FaceX
No description available
Ask AI about FaceX
Powered by Claude · Grounded in docs
I know everything about FaceX. Ask me about installation, configuration, usage, or troubleshooting.
0/500
Reviews
Documentation
Shufti Pro MCP Server / Shufti Pro MCP 服务器
English
Overview
This project is a standalone Model Context Protocol (MCP) server dedicated to Shufti Pro identity verification workflows.
It is designed for premium compliance scenarios that require strong liveness checks and high-quality document OCR extraction.
Server
- Name:
mcp-server-shuftipro - Entry file:
server.ts - Transport: MCP stdio
Exposed Tools
shufti_face_liveness_match
- Purpose: 1:1 face match with liveness detection.
- Best for: biometric login and step-up authentication.
- Input:
selfie_base64(string, required)document_face_base64(string, required)reference_id(string, required)
shufti_full_document_ocr
- Purpose: OCR extraction (including MRZ/text) and document authenticity checks.
- Best for: initial KYC data capture.
- Input:
document_base64(string, required)document_type(string, required, e.g.passport,id_card)reference_id(string, required)
Unified Output Schema
Both tools return the same structured output:
{
"status": "success | failed | error",
"event": "verification.accepted | verification.declined | ...",
"decline_reason": "string | null",
"extracted_data": "object | null"
}
Environment Variables
Set these before running:
SHUFTI_CLIENT_ID(required)SHUFTI_SECRET_KEY(required)SHUFTI_API_URL(optional, default:https://api.shuftipro.com/)
Installation
Install dependencies in your TypeScript MCP project:
npm install @modelcontextprotocol/sdk axios
Run
Build and run with your project setup (example):
npx tsc
node dist/server.js
Notes
- The server validates base64 inputs before sending API requests.
- The server uses Basic Auth generated from
SHUFTI_CLIENT_IDandSHUFTI_SECRET_KEY. - Errors are normalized into the same output schema for stable orchestrator parsing.
中文
项目简介
本项目是一个专用于 Shufti Pro 身份核验流程的独立 MCP(Model Context Protocol)服务器。
它面向高合规场景,支持高质量活体检测与证件 OCR 数据提取。
服务信息
- 服务名:
mcp-server-shuftipro - 入口文件:
server.ts - 通信方式:MCP stdio
暴露工具
shufti_face_liveness_match
- 作用:执行 1:1 人脸比对与活体检测。
- 适用场景:生物识别登录、二次校验(step-up)。
- 输入参数:
selfie_base64(字符串,必填)document_face_base64(字符串,必填)reference_id(字符串,必填)
shufti_full_document_ocr
- 作用:提取证件 OCR(含 MRZ/文本)并进行证件真实性校验。
- 适用场景:初次 KYC 信息采集。
- 输入参数:
document_base64(字符串,必填)document_type(字符串,必填,例如passport、id_card)reference_id(字符串,必填)
统一输出结构
两个工具都返回统一结构:
{
"status": "success | failed | error",
"event": "verification.accepted | verification.declined | ...",
"decline_reason": "string | null",
"extracted_data": "object | null"
}
环境变量
运行前请设置:
SHUFTI_CLIENT_ID(必填)SHUFTI_SECRET_KEY(必填)SHUFTI_API_URL(可选,默认:https://api.shuftipro.com/)
安装依赖
在你的 TypeScript MCP 项目中安装依赖:
npm install @modelcontextprotocol/sdk axios
运行方式
按你的项目构建流程运行(示例):
npx tsc
node dist/server.js
说明
- 服务会在调用外部 API 前校验 base64 输入。
- 服务通过
SHUFTI_CLIENT_ID与SHUFTI_SECRET_KEY生成 Basic Auth。 - 所有异常会归一化为统一输出结构,方便上游编排器稳定解析。
