Skip to main content
Glama
enesjakozh
Cognitive-Stack

Orion Vision MCP Server

by Cognitive-Stack
OverviewSchemaSourceRelated ServersDiscussionsAdmin
TypeScript
Remote

오리온 비전 MCP 서버 🚀

🔌 Cline, Cursor, Claude Desktop 및 기타 MCP 클라이언트와 호환됩니다!

Orion Vision MCP는 모든 MCP 클라이언트와도 호환됩니다.

MCP(Model Context Protocol)는 AI 시스템이 다양한 데이터 소스 및 도구와 원활하게 상호 작용할 수 있도록 하는 개방형 표준으로, 안전한 양방향 연결을 용이하게 합니다.

Orion Vision MCP 서버는 다음을 제공합니다.

  • Azure Form Recognizer/Document Intelligence와의 원활한 통합

  • 문서 분석 및 양식 데이터 추출 기능

  • 다양한 문서 유형(영수증, 송장, 신분증 등) 지원

  • TypeScript를 사용한 유형 안전 작업

필수 조건 🔧

시작하기 전에 다음 사항을 확인하세요.

  • Azure Form Recognizer/Document Intelligence 엔드포인트 및 키

  • 클로드 데스크탑 또는 커서

  • Node.js(v20 이상)

  • Git 설치됨(Git 설치 방법을 사용하는 경우에만 필요)

Related MCP server: pdf-mcp

Orion Vision MCP 서버 설치 ⚡

NPX로 실행

지엑스피1

Smithery를 통해 설치

Smithery를 통해 Claude Desktop에 Orion Vision MCP 서버를 자동으로 설치하려면:

npx -y @smithery/cli install @orion-vision/mcp --client claude

MCP 클라이언트 구성 ⚙️

Cline 구성 🤖

클라인에 Orion Vision MCP 서버를 설정하는 가장 쉬운 방법은 마켓플레이스를 통해 한 번의 클릭으로 설정하는 것입니다.

  1. VS Code에서 Cline 열기

  2. 사이드바에서 Cline 아이콘을 클릭하세요

  3. "MCP 서버" 탭(4개 사각형)으로 이동합니다.

  4. "Orion Vision"을 검색하고 "설치"를 클릭하세요.

  5. 메시지가 표시되면 Azure Form Recognizer 자격 증명을 입력하세요.

또는 Cline에서 Orion Vision MCP 서버를 수동으로 설정할 수 있습니다.

  1. Cline MCP 설정 파일을 엽니다.

# For macOS:
code ~/Library/Application\ Support/Code/User/globalStorage/saoudrizwan.claude-dev/settings/cline_mcp_settings.json

# For Windows:
code %APPDATA%\Code\User\globalStorage\saoudrizwan.claude-dev\settings\cline_mcp_settings.json

  1. Orion Vision 서버 구성을 파일에 추가합니다.

{
  "mcpServers": {
    "orion-vision-mcp": {
      "command": "npx",
      "args": ["-y", "orion-vision-mcp@latest"],
      "env": {
        "AZURE_FORM_RECOGNIZER_ENDPOINT": "your-endpoint-here",
        "AZURE_FORM_RECOGNIZER_KEY": "your-key-here"
      },
      "disabled": false,
      "autoApprove": []
    }
  }
}

  1. 파일을 저장하고 Cline이 이미 실행 중이면 다시 시작합니다.

커서 구성 🖥️

참고 : 커서 버전 0.45.6 이상이 필요합니다.

Cursor에서 Orion Vision MCP 서버를 설정하려면:

  1. 커서 설정 열기

  2. 기능 > MCP 서버로 이동

  3. "+ 새 MCP 서버 추가" 버튼을 클릭하세요

  4. 다음 정보를 입력하세요:

    • 이름 : 서버의 별명을 입력하세요(예: "orion-vision-mcp")

    • 유형 : 유형으로 "명령"을 선택하세요

    • 명령어 : 서버를 실행하기 위한 명령어를 입력하세요: GXP5

    중요 : your-endpoint 및 your-key Azure Form Recognizer 자격 증명으로 바꾸세요.

Claude Desktop 앱 구성하기 🖥️

macOS의 경우:

# Create the config file if it doesn't exist
touch "$HOME/Library/Application Support/Claude/claude_desktop_config.json"

# Opens the config file in TextEdit
open -e "$HOME/Library/Application Support/Claude/claude_desktop_config.json"

Windows의 경우:

code %APPDATA%\Claude\claude_desktop_config.json

Orion Vision 서버 구성을 추가합니다.

{
  "mcpServers": {
    "orion-vision-mcp": {
      "command": "npx",
      "args": ["-y", "orion-vision-mcp@latest"],
      "env": {
        "AZURE_FORM_RECOGNIZER_ENDPOINT": "your-endpoint-here",
        "AZURE_FORM_RECOGNIZER_KEY": "your-key-here"
      }
    }
  }
}

Claude 데스크톱 앱에서의 사용법 🎯

설치가 완료되고 Claude 데스크톱 앱 구성이 완료되면 Claude 데스크톱 앱을 완전히 닫았다가 다시 열어야 orion-vision-mcp 서버를 확인할 수 있습니다. 앱 왼쪽 하단에 사용 가능한 MCP 도구를 나타내는 망치 아이콘이 표시됩니다.

오리온 비전 예시

  1. 문서 분석 :

Analyze the document at "https://example.com/document.pdf" using Azure Form Recognizer.

  1. 양식 데이터 추출 :

Extract data from the invoice at "https://example.com/invoice.pdf".

  1. 프로세스 ID 문서 :

Process the ID document at "https://example.com/id.pdf" and extract relevant information.

문제 해결 🛠️

일반적인 문제

  1. 서버를 찾을 수 없습니다

    • npm --version 실행하여 npm 설치를 확인하세요.

    • Claude Desktop 구성 구문 확인

    • node --version 실행하여 Node.js가 제대로 설치되었는지 확인하세요.

  2. Azure Form Recognizer 자격 증명 문제

    • Azure Form Recognizer 엔드포인트와 키가 유효한지 확인하세요.

    • 구성에서 자격 증명이 올바르게 설정되었는지 확인하세요.

    • 자격 증명 주위에 공백이나 따옴표가 없는지 확인하세요.

  3. 문서 처리 문제

    • 문서 URL에 접근할 수 있는지 확인하세요

    • 문서 형식이 지원되는지 확인하세요

    • 문서가 손상되었거나 암호로 보호되지 않았는지 확인하세요.

감사의 말 ✨

  • MCP 사양을 위한 모델 컨텍스트 프로토콜

  • 클로드 데스크탑을 위한 인류학

  • Microsoft Azure for Form Recognizer/Document Intelligence

Available Tools

2 tools
analyze-documentC

Analyzes a document using Azure Form Recognizer and returns structured data

ParametersJSON Schema
NameRequiredDescriptionDefault
modelIdNoOptional model ID for custom models
urlYesURL of the document to analyze

TDQS

C2.9/5.0
Behavior2/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations provided, the description carries the full burden of behavioral disclosure. It mentions the analysis method and output type but omits critical details like authentication needs, rate limits, processing time, error handling, or what 'structured data' entails, leaving significant gaps for a tool performing external API calls.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is a single, efficient sentence with zero waste, clearly front-loading the core functionality. Every word contributes to understanding the tool's purpose without redundancy.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness2/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the complexity of document analysis with an external service, no annotations, and no output schema, the description is incomplete. It lacks details on behavioral traits, output format, error cases, and usage context, which are essential for effective tool invocation.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100%, so the schema already documents both parameters (modelId and url). The description adds no additional parameter semantics beyond what the schema provides, such as examples or constraints, meeting the baseline for high coverage.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the action ('analyzes') and resource ('a document') using Azure Form Recognizer, with the outcome of returning structured data. It distinguishes from the sibling 'extract-form-data' by specifying the analysis method, though not explicitly contrasting their use cases.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines2/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

No guidance is provided on when to use this tool versus the sibling 'extract-form-data' or other alternatives. The description implies usage for document analysis but lacks context on prerequisites, constraints, or comparative scenarios.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

extract-form-dataC

Extracts structured data from forms using Azure Form Recognizer

ParametersJSON Schema
NameRequiredDescriptionDefault
formTypeYesType of form to analyze
urlYesURL of the form document to analyze

TDQS

C2.9/5.0
Behavior2/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

No annotations are provided, so the description carries the full burden of behavioral disclosure. While it mentions the technology (Azure Form Recognizer), it doesn't describe what happens during extraction - whether it's a read-only operation, if it modifies data, authentication requirements, rate limits, or error handling. For a tool with no annotation coverage, this leaves significant behavioral gaps.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is extremely concise - a single sentence that directly states the tool's purpose without any unnecessary words. It's front-loaded with the core functionality and doesn't waste space on redundant information. Every word earns its place in this minimal description.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness2/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given that there's no output schema and no annotations, the description should provide more context about what the tool returns and how it behaves. For a data extraction tool with 2 required parameters, the description is too minimal - it doesn't explain the extraction results format, error conditions, or practical usage examples. The completeness is inadequate for the tool's complexity.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100%, so the schema already documents both parameters thoroughly. The description doesn't add any meaningful parameter semantics beyond what's in the schema - it doesn't explain how 'formType' affects extraction results or provide examples of valid URLs. With complete schema coverage, the baseline score of 3 is appropriate.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool's purpose: 'Extracts structured data from forms using Azure Form Recognizer'. It specifies the action (extracts), resource (structured data from forms), and technology (Azure Form Recognizer). However, it doesn't explicitly differentiate from its sibling 'analyze-document', which might have overlapping functionality.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines2/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description provides no guidance on when to use this tool versus its sibling 'analyze-document' or other alternatives. It doesn't mention prerequisites, limitations, or specific scenarios where this tool is preferred. The only implied context is that it works with forms, but no explicit usage instructions are given.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

  1. 2 tool updatesv1.0.0
    • First observedanalyze-document
    • First observedextract-form-data
Deploy Server
One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

TDQS

C2.6/5.0

Scored across 2 tools

Disambiguation1/5

The two tools have nearly identical purposes: both use Azure Form Recognizer to extract structured data from documents/forms. 'analyze-document' and 'extract-form-data' are functionally indistinguishable, with no clear boundary between them. This high ambiguity will cause agents to misselect between tools.

Naming Consistency3/5

Both tools use kebab-case naming, which is consistent. However, the verb choices ('analyze' vs 'extract') are different despite similar functionality, creating minor inconsistency. The naming pattern is readable but not perfectly aligned in purpose.

Tool Count2/5

With only 2 tools, the server feels thin for a vision/document processing domain. A typical MCP server for this scope would include more operations like text extraction, image analysis, or OCR configuration. The minimal tool count limits functionality and suggests incomplete coverage.

Completeness2/5

For a vision/document processing server, there are significant gaps: no image analysis tools, no OCR configuration, no batch processing, and no support for different document types beyond forms. The surface is severely limited, focusing only on form data extraction without broader vision capabilities.

Maintenance

ActivityInactive
ResponsivenessNo issues

Resources

Related MCP Connectors

Related MCP Servers

Appeared in Searches