sifter-mcp
OfficialSifter
结构化任何文档。像查询数据库一样查询它。通过 API 构建应用。
开源文档智能引擎 — 基于模式的提取、自然语言查询、MCP 服务器、Python 和 TypeScript SDK。基于 MIT 协议可自托管。

为什么不用 RAG?
RAG 是为检索而构建的 — 帮我找到与此查询相似的片段。当面对发票、合同或收据等同质化集合时,它会失效,因为这些文档看起来都一样,而问题通常是聚合分析,而非相似度搜索。

Sifter 的方法:一次性提取结构化字段(客户、日期、总额),将它们存储为类型化记录,并使用真实的过滤器和聚合进行查询。答案是精确且可复现的 — 因为它是数据库查询,而不是相似度搜索。
Related MCP server: API Agent
快速入门
git clone https://github.com/sifter-ai/sifter
cd sifter/code
cp server/.env.example server/.env.local # set SIFTER_DEFAULT_API_KEY (required)
docker compose up -d打开 http://localhost:3000 — 创建一个 sift,上传文档,查询结果。
Python SDK
pip install sifter-aifrom sifter import Sifter
s = Sifter(api_key="sk-...")
sift = s.create_sift("Invoices", "client name, date, total amount")
sift.upload("./invoices/")
sift.wait()
for record in sift.records():
print(record["extracted_data"])
# {"client": "Acme Corp", "date": "2024-01-15", "total_amount": 1500.0}TypeScript SDK
npm install @sifter-ai/sdkimport { Sifter } from "@sifter-ai/sdk";
const client = new Sifter({ apiKey: "sk-..." });
const sift = await client.createSift("Invoices", "client, date, total amount");
await sift.upload("./invoices/");
await sift.wait();
const records = await sift.records();
console.log(records);MCP 服务器 (Claude Desktop / Cursor / AI 代理)
{
"mcpServers": {
"sifter": {
"command": "uvx",
"args": ["sifter-mcp", "--base-url", "http://localhost:8000"],
"env": { "SIFTER_API_KEY": "sk-dev" }
}
}
}然后问 Claude:“上个季度所有发票的未付款总额是多少?”
想要一个无需运行本地服务器的远程 MCP URL?→ Sifter Cloud
功能特性
基于模式的提取 — 用自然语言描述要提取的内容;模式会自动推断并导出为 Pydantic / TypeScript 类型
自然语言查询 — 用普通语言提问;Sifter 会生成可检查的 MongoDB 聚合管道
MCP 服务器 — stdio 传输,读写工具,无需自定义集成代码
REST API + SDKs — 完整的 OpenAPI 规范,提供 Python 和 TypeScript 的类型化客户端
Webhooks — 每个提取事件触发 HMAC 签名的 HTTP 回调
基于规范的仪表板 — 短自然语言规范 → 自动生成看板(KPI、细分、表格、时间序列)
CLI —
sifter extract、sifter records、sifter sifts,适用于终端工作流和 CI可自托管 — Docker Compose,自带 MongoDB 和 LLM API 密钥
不想运行基础设施?
Sifter Cloud 是托管版本 — 无需 Mongo,无需运维,提供远程 MCP 端点,支持 Google Drive 和电子邮件接入。提供免费层级。
文档
完整文档请访问 docs.sifter.run — 包含快速入门、SDK 参考、MCP 指南、操作手册和自托管说明。
许可证
MIT — 详见 LICENSE。
由 Bruno Fortunato 创建。
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Related MCP Servers
- FlicenseNot gradedqualityDmaintenanceTransforms PDF collections into a searchable knowledge base using TF-IDF indexing and proximity matching. It enables users to search documents, retrieve specific page content, and manage document libraries through natural language via MCP clients.5
- AlicenseNot gradedqualityDmaintenanceTurn any API into an MCP server. Query in English. Get results—even when the API can't.284MIT

flexorch-mcpofficial
AlicenseAqualityAmaintenanceEnables Claude and other MCP-compatible agents to process documents, extract structured data, detect PII, and export LLM-ready datasets through natural language tool calls.81MIT- AlicenseAqualityBmaintenanceA local-first MCP server that ingests PDFs, extracts structure, and provides semantic search and sequential navigation tools for AI clients to query and learn from documents.10MIT
Related MCP Connectors
Analytical memory for AI agents: a real Postgres queried in plain English over MCP. One command.
Turn any PDF into structured JSON via AI + OCR: invoices, bank statements, contracts.
GibsonAI MCP server: manage your databases with natural language
Latest Blog Posts
- Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)By Om-Shree-0709 on .mcpAgent IdentityOAuth 2.1
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/sifter-ai/sifter'
If you have feedback or need assistance with the MCP directory API, please join our Discord server