Unstructured UNS MCP

왜 쓰나요

핵심 기능

60+ source connectors (S3, Drive, OneDrive, SharePoint, GCS)
Document partitioning into elements (Title, Table, NarrativeText, ListItem)
Built-in chunkers (by_title, basic)
Embeddings via OpenAI, Voyage, or local models
Targets: vector DBs (Pinecone, Weaviate, Qdrant) or warehouses

라이브 데모

실제 사용 모습

unstructured-uns-mcp.replay ▶ 준비됨

0/0

설치

클라이언트 선택

~/Library/Application Support/Claude/claude_desktop_config.json · Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "unstructured-uns-mcp": {
      "command": "uvx",
      "args": [
        "uns-mcp"
      ]
    }
  }
}

Claude Desktop → Settings → Developer → Edit Config 열기. 저장 후 앱 재시작.

~/.cursor/mcp.json · .cursor/mcp.json

{
  "mcpServers": {
    "unstructured-uns-mcp": {
      "command": "uvx",
      "args": [
        "uns-mcp"
      ]
    }
  }
}

Cursor는 Claude Desktop과 동일한 mcpServers 스키마 사용. 프로젝트 설정이 전역보다 우선.

VS Code → Cline → MCP Servers → Edit

{
  "mcpServers": {
    "unstructured-uns-mcp": {
      "command": "uvx",
      "args": [
        "uns-mcp"
      ]
    }
  }
}

Cline 사이드바의 MCP Servers 아이콘 클릭 후 "Edit Configuration" 선택.

~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "unstructured-uns-mcp": {
      "command": "uvx",
      "args": [
        "uns-mcp"
      ]
    }
  }
}

Claude Desktop과 같은 형식. Windsurf 재시작 후 적용.

~/.continue/config.json

{
  "mcpServers": [
    {
      "name": "unstructured-uns-mcp",
      "command": "uvx",
      "args": [
        "uns-mcp"
      ]
    }
  ]
}

Continue는 맵이 아닌 서버 오브젝트 배열 사용.

~/.config/zed/settings.json

{
  "context_servers": {
    "unstructured-uns-mcp": {
      "command": {
        "path": "uvx",
        "args": [
          "uns-mcp"
        ]
      }
    }
  }
}

context_servers에 추가. 저장 시 Zed가 핫 리로드.

claude mcp add unstructured-uns-mcp -- uvx uns-mcp

한 줄 명령. claude mcp list로 확인, claude mcp remove로 제거.

사용 사례

실전 활용법: Unstructured UNS MCP

Stand up a RAG pipeline from a SharePoint folder to Pinecone

👤 Data engineers / RAG builders ⏱ ~15 min intermediate

언제 쓸까: You have a corporate doc dump and want a Claude-queryable index.

사전 조건

Server/skill installed and authenticated — See repo README

흐름

Define the pipeline

Create an Unstructured pipeline: source SharePoint folder X, partition by_title with 1024 token max, embed with text-embedding-3-small, target Pinecone index 'corp-docs'.✓ 복사됨

→ Pipeline id
Run and monitor

Run it and tell me when it's done. Report any failed documents.✓ 복사됨

→ Status updates + final summary

결과: Production-grade ingest with proper chunking — not naive PDF text dumps.

함정

Default chunkers can split tables across chunks. For dense tabular docs, use 'by_title' with combine_text_under_n_chars. — Default chunkers can split tables across chunks. For dense tabular docs, use 'by_title' with combine_text_under_n_chars.

함께 쓰기: filesystem · qdrant-mcp-server

조합

다른 MCP와 조합해 10배 효율

unstructured-uns-mcp + filesystem

Pair with filesystem for complementary capabilities

Use this server together with filesystem to complete a multi-step task.✓ 복사됨

unstructured-uns-mcp + qdrant-mcp-server

Pair with qdrant-mcp-server for complementary capabilities

Use this server together with qdrant-mcp-server to complete a multi-step task.✓ 복사됨

도구

이 MCP가 노출하는 것

도구	입력	언제 호출	비용
create_pipeline	source, dest, partition_args, chunk_args	Define a new ingest job	1 API call
run_pipeline	pipeline_id	Execute the pipeline	Per Unstructured plan
list_workflows	(none)	See all configured workflows	1 API call

비용 및 제한

운영 비용

API 쿼터: See provider docs for rate limits
호출당 토큰: Varies by tool
금액: See repo README for pricing details
팁: Cache tool results and avoid repeated identical calls.

보안

권한, 시크릿, 파급범위

자격 증명 저장: Use environment variables; never commit secrets

데이터 외부 송신: Tool calls go to the provider's API as documented

문제 해결

자주 발생하는 오류와 해결

401 Unauthorized

Get an API key at unstructured.io → Settings → API keys; set UNSTRUCTURED_API_KEY and UNSTRUCTURED_API_URL.

확인: list_workflows returns at least one

source connector auth failure

Connector creds (e.g. SharePoint app token) are configured per-source. Recreate the source via Unstructured UI to refresh.

확인: Run a 1-file test pipeline first

대안