Skip to main content
Glama
1617110693

Image Parse MCP

by 1617110693
OverviewSchemaRelated ServersScoreDiscussions
Python
Remote

Image Parse MCP

中文文档

A multimodal image analysis MCP server that connects to any OpenAI-compatible vision API. Provide an image (URL, local path, or base64) and a prompt — get back a detailed analysis from the multimodal LLM of your choice.

Supported Providers

Any provider with an OpenAI-compatible chat completions endpoint:

  • OpenAI — GPT-4o, GPT-4-vision, GPT-4.1-mini

  • Anthropic (via compatible proxy / gateway)

  • Google Gemini (via OpenAI-compatible endpoint)

  • Azure OpenAI

  • Local models (Ollama, vLLM, LM Studio with OpenAI-compatible servers)

  • Alibaba Bailian — Qwen-VL (via OpenAI-compatible endpoint)

  • Third-party (DeepSeek, Groq, Together.ai, OpenRouter, etc.)

Related MCP server: Vision MCP Server

Configuration

Set these environment variables before launching the server:

Variable

Required

Default

Description

IMAGE_PARSE_API_KEY

Yes

—

API key for your provider

IMAGE_PARSE_BASE_URL

No

https://api.openai.com/v1

API base URL

IMAGE_PARSE_MODEL

No

gpt-4o

Multimodal model name

Example: OpenAI

export IMAGE_PARSE_API_KEY=sk-...
export IMAGE_PARSE_BASE_URL=https://api.openai.com/v1
export IMAGE_PARSE_MODEL=gpt-4o

Example: Google Gemini (via AI Studio)

export IMAGE_PARSE_API_KEY=your-gemini-api-key
export IMAGE_PARSE_BASE_URL=https://generativelanguage.googleapis.com/v1beta/openai
export IMAGE_PARSE_MODEL=gemini-2.5-flash

Example: Ollama (local)

export IMAGE_PARSE_API_KEY=ollama
export IMAGE_PARSE_BASE_URL=http://localhost:11434/v1
export IMAGE_PARSE_MODEL=llava

Example: Azure OpenAI

export IMAGE_PARSE_API_KEY=your-azure-api-key
export IMAGE_PARSE_BASE_URL=https://your-resource.openai.azure.com/openai/deployments/your-deployment
export IMAGE_PARSE_MODEL=gpt-4o

Example: Alibaba Bailian (Qwen-VL)

export IMAGE_PARSE_API_KEY=your-dashscope-api-key
export IMAGE_PARSE_BASE_URL=https://dashscope.aliyuncs.com/compatible-mode/v1
export IMAGE_PARSE_MODEL=qwen-vl-max

Example: DeepSeek

export IMAGE_PARSE_API_KEY=your-deepseek-api-key
export IMAGE_PARSE_BASE_URL=https://api.deepseek.com/v1
export IMAGE_PARSE_MODEL=deepseek-chat

Install & Run

# Clone or enter the project directory
cd image-parse

# Run directly (uv handles venv + deps automatically)
uv run image-parse-mcp

Claude Code Configuration

Add to your Claude Code MCP config (~/.claude/claude.json or project .mcp.json):

{
  "mcpServers": {
    "image-parse": {
      "type": "stdio",
      "command": "uv",
      "args": ["run", "--directory", "path/to/image-parse", "image-parse-mcp"],
      "env": {
        "IMAGE_PARSE_API_KEY": "sk-...",
        "IMAGE_PARSE_BASE_URL": "https://api.openai.com/v1",
        "IMAGE_PARSE_MODEL": "gpt-4o"
      }
    }
  }
}

Tool: analyze_image

Parameter

Required

Description

image_source

Yes

URL, local file path, base64 string, or data URI

prompt

Yes

What to analyze / extract from the image

mime_type

No

Override auto-detected MIME type (e.g. image/webp)

What agents use it for

  • Describe the contents of an image

  • Extract text from a screenshot (OCR)

  • Read and interpret charts, graphs, data visualizations

  • Analyze UI screenshots (layout, elements, issues)

  • Identify objects, colours, people, or scenes in photos

  • Compare visual information across multiple images

  • Diagnose errors from error screenshots

Input forms for image_source

# URL
https://example.com/screenshot.png

# Local file path (on the host machine)
/Users/me/Downloads/chart.png

# Base64 data URI
data:image/png;base64,iVBORw0KGgo...

# Raw base64
iVBORw0KGgo...

Development

# Create venv and install deps
uv venv
uv pip install -e .

# Run tests
uv run python -m pytest

Available Tools

1 tool
analyze_imageA
Read-onlyIdempotent

Analyze an image with a multimodal LLM (GPT-4o, Claude, Gemini, etc.).

Provide an image (URL, local path, or base64) and a description of what you want to know. The tool calls an OpenAI-compatible vision API and returns the model's text response.

Use this tool whenever you have an image and need to:

  • Describe its contents

  • Extract text / OCR

  • Understand a chart, diagram, or data visualization

  • Analyse a UI screenshot (layout, elements, issues)

  • Identify objects, colours, people, or scenes in a photo

  • Compare or summarise visual information

Args: params (AnalyzeImageInput): - image_source (str): URL, local file path, or base64 image data. - prompt (str): What to analyze or extract from the image. - mime_type (Optional[str]): Override auto-detected MIME type.

Returns: str: The multimodal model's analysis as plain text / Markdown.

ParametersJSON Schema
NameRequiredDescriptionDefault
paramsYes

Output Schema

ParametersJSON Schema
NameRequiredDescription
resultYes

TDQS

A4.4/5.0
Behavior4/5

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

Annotations already provide readOnlyHint=true, destructiveHint=false, idempotentHint=true. The description adds behavioral context: it calls an OpenAI-compatible vision API, accepts various image input forms (URL, local path, base64), and returns text/Markdown. This adds value beyond annotations without contradiction.

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

Conciseness4/5

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

The description is well-structured: a clear opening sentence, a bulleted list of use cases, and a section explaining arguments. It is mostly concise, though some examples could be trimmed. It front-loads the most important information.

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

Completeness5/5

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

The description covers all essential aspects: input formats, prompt style, optional MIME override, and return type. Given that an output schema exists, the description effectively complements it by explaining the return value as 'plain text / Markdown.' This provides a complete picture.

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

Parameters4/5

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

Although the schema has detailed descriptions for each parameter (image_source, prompt, mime_type), the tool description adds extra context with examples and guidance on prompt specificity. This enhances the schema's information, earning a score above baseline.

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

Purpose5/5

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

The description clearly states that the tool analyzes an image using a multimodal LLM, and lists specific use cases (describe contents, OCR, charts, UI screenshots, etc.). It is specific about the action and resource, and distinguishes itself from siblings (none listed).

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

Usage Guidelines4/5

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

The description provides clear context by listing when to use the tool (e.g., whenever you have an image and need to describe, extract text, analyze charts). It does not explicitly state when not to use or mention alternatives, but given the absence of sibling tools, this is acceptable.

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

Install Server
RayrunBuild, deploy, and operate MCP servers without becoming an infrastructure team.ray.runAd

TDQS

A4.4/5.0
Disambiguation5/5

With only one tool, there is no possibility of confusion or overlap. Agents can unambiguously select the correct tool for any image analysis task.

Naming Consistency5/5

The single tool uses a clear verb_noun pattern ('analyze_image'), which is consistent with best practices. No other tools exist to introduce inconsistency.

Tool Count4/5

One tool is on the low end for a general utility, but it fits the narrow scope of image parsing via LLM. A single unified tool can be sufficient if it covers the intended use cases without needing multiple specialized tools.

Completeness4/5

The tool covers core image analysis needs (description, OCR, chart understanding) through a multimodal LLM. Minor gaps like batch processing or model selection are absent, but the core functionality is well-covered for the stated purpose.

Maintenance

ActivityStale
ResponsivenessNo issues

Resources

Unclaimed servers have limited discoverability.

Looking for Admin?

If you are the server author, to access and configure the admin panel.

Related MCP Connectors

Related MCP Servers

Latest Blog Posts

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/1617110693/Image-Parse-MCP'

If you have feedback or need assistance with the MCP directory API, please join our Discord server