How do I use vision-mcp?

1. Click on "Deploy Server". 2. Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state. 3. In the chat, type @ followed by the MCP server name and your instructions, e.g., "@vision-mcp Extract text from image at /tmp/screenshot.png" That's it! The server will respond to your query, and you can continue using it as needed. Here is a step-by-step guide with screenshots.

vision-mcp

by 666666999999666

Overview Schema Source Related Servers Discussions Admin

Python

Hybrid

vision-mcp

English | 中文

Give MCP-compatible AI agents image analysis, metadata inspection, cropping, OCR, and image comparison through any OpenAI-compatible vision model.

Features

Analyze screenshots, charts, documents, UI, objects, and general images.
Inspect image dimensions and metadata without calling a model.
Crop and zoom into regions using normalized coordinates.
Extract visible text with a VLM or an optional dedicated OCR model.
Compare two to four images.
Accept public URLs, local files, data URLs, and Base64 images.
Run locally over the standard MCP stdio transport.

Related MCP server: vision-bridge-mcp

Claude Code

Requirements

Python 3.10 or newer
uv
An OpenAI-compatible vision API endpoint and API key

uvx downloads the published package from PyPI into an isolated environment and runs it. It does not use the source code in your current directory and does not permanently install the package into your system Python.

Add To Claude Code

The command below configures Claude Code to start vision-mcp from PyPI:

claude mcp add --scope user vision \
  --env UV_DEFAULT_INDEX=https://pypi.org/simple \
  VISION_API_KEY="your-api-key" \
  VISION_BASE_URL="https://your-provider.example/v1" \
  VISION_MODEL_ID="your-vision-model" \
  -- uvx vision-mcp

Use UV_DEFAULT_INDEX=https://pypi.org/simple when your local PyPI mirror has not synchronized the latest release.

Verify the connection:

claude mcp get vision
claude mcp list

Then start Claude Code and ask:

Use vision_capabilities to show the available vision tools.

Analyze a local image:

Use vision_inspect on /data/example.png, then use vision_analyze to describe it.

By default, local image access is limited to /data and /tmp. Add another directory with:

claude mcp remove --scope user vision

claude mcp add --scope user vision \
  --env UV_DEFAULT_INDEX=https://pypi.org/simple \
  VISION_API_KEY="your-api-key" \
  VISION_BASE_URL="https://your-provider.example/v1" \
  VISION_MODEL_ID="your-vision-model" \
  VISION_ALLOWED_PATHS="/data,/tmp,/home/your-user/Pictures" \
  -- uvx vision-mcp

Dedicated OCR Model

Without dedicated OCR configuration, vision_extract_text uses the configured vision model. To use a separate OCR model:

claude mcp add --scope user vision \
  --env UV_DEFAULT_INDEX=https://pypi.org/simple \
  VISION_API_KEY="your-vision-api-key" \
  VISION_BASE_URL="https://your-provider.example/v1" \
  VISION_MODEL_ID="your-vision-model" \
  OCR_ENABLED=true \
  OCR_API_KEY="your-ocr-api-key" \
  OCR_BASE_URL="https://your-provider.example/v1" \
  OCR_MODEL_ID="your-ocr-model" \
  -- uvx vision-mcp

Never commit real API keys to Git.

Other MCP Clients

Use this stdio configuration with MCP clients that accept JSON configuration:

{
  "mcpServers": {
    "vision": {
      "command": "uvx",
      "args": ["vision-mcp"],
      "env": {
        "UV_DEFAULT_INDEX": "https://pypi.org/simple",
        "VISION_API_KEY": "your-api-key",
        "VISION_BASE_URL": "https://your-provider.example/v1",
        "VISION_MODEL_ID": "your-vision-model"
      }
    }
  }
}

Tools

Tool	Purpose
`vision_analyze`	Analyze an image with task-specific prompts
`vision_inspect`	Read image dimensions, format, size, and mode
`vision_crop_analyze`	Crop and analyze a normalized image region
`vision_extract_text`	Extract visible text using OCR or the VLM
`vision_compare`	Compare two to four images
`vision_capabilities`	Show server configuration and limits

Response format

Every tool returns a JSON string. Clients must json.loads the result before reading any field. All top-level keys are always present (even when empty), so consumers can iterate the envelope without dict.get(...) guards.

Success envelope

{
  "schema_version": "1.0",
  "ok": true,
  "tool": "vision_analyze",
  "task": "general",
  "model": "...",
  "source": null,
  "sources": [],
  "result": {},
  "warnings": [],
  "raw_model_output": null,
  "error": null
}

Field	Type	When set
`schema_version`	`string`	Always. Currently `"1.0"`.
`ok`	`bool`	Always. `true` on success, `false` on failure.
`tool`	`string`	Always. The tool name (e.g. `vision_analyze`).
`task`	`string \| null`	The `task` argument when the tool takes one; `null` for `vision_capabilities` and `vision_extract_text`.
`model`	`string \| null`	The configured model identifier (e.g. `glm-4v-flash`). Set even on failure when the tool knew it.
`source`	`SourceMeta \| null`	Single-image tools. `null` for `vision_compare` and `vision_capabilities`.
`sources`	`SourceMeta[]`	`vision_compare` only: one entry per input image. Empty for all other tools.
`result`	`object`	Tool-specific (see below). `null` on failure.
`warnings`	`string[]`	Always a list (empty on success). Soft-failure notes (e.g. `vision_extract_text` falling back from OCR to VLM).
`raw_model_output`	`object \| null`	Sanitized provider response when `include_raw=true`; `null` otherwise.
`error`	`ErrorPayload \| null`	`null` on success. Populated on failure.

SourceMeta fields: type (url / file / data_url / base64), mime_type, width, height, size_bytes, source_ref (only when include_source_ref=true; redacted to host/path for URLs or basename for files; null for data URLs and base64).

Failure envelope

{
  "schema_version": "1.0",
  "ok": false,
  "tool": "vision_analyze",
  "task": "general",
  "model": "...",
  "source": null,
  "sources": [],
  "result": null,
  "warnings": [],
  "raw_model_output": null,
  "error": {
    "code": "INVALID_INPUT",
    "message": "Input is not a valid supported image",
    "retryable": false,
    "details": {}
  }
}

error.code values: INVALID_INPUT, IMAGE_TOO_LARGE, UNSUPPORTED_FORMAT, SECURITY_ERROR, PROVIDER_ERROR, TIMEOUT, INTERNAL_ERROR. retryable=true means the caller may try the same call again.

Per-tool `result` shape

Tool	`result` keys
`vision_analyze`	`summary`, `observations[]`, `inferences[]`, `uncertainties[]`, `suggested_followups[]`
`vision_extract_text`	`text`, `blocks[]`, `layout_preserved`, `unclear_segments[]`
`vision_compare`	`summary`, `differences[]`, `same_elements[]`
`vision_crop_analyze`	`crop: {x, y, width, height}`, `summary`, `observations[]`
`vision_inspect`	`width`, `height`, `format`, `mime_type`, `mode`, `size_bytes`, `has_transparency`, `source_type`
`vision_capabilities`	`server`, `version`, `vlm_provider`, `ocr_provider`, `ocr_enabled`, `tools`, `supports`, `limits`, `task_types`

Arrays that are not yet parsed from model output are returned as empty arrays (no fabricated structure). observations, inferences, and differences are empty in the current release; only summary carries the model's free-form text.

Multi-image input

vision_compare accepts 2–4 images. The envelope reports them in sources: [SourceMeta, ...] (one entry per input, in input order). source is null for multi-image tools. All other image tools accept a single image and use source; sources is [].

Opt-in flags

include_raw: bool = False — when true, raw_model_output contains a sanitized subset of the provider response: {model, response_metadata: {model_name, finish_reason, system_fingerprint}, usage_metadata: {input_tokens, output_tokens, total_tokens}}. HTTP headers, request IDs, signed URLs, and raw exception text are dropped before reaching the envelope. Off by default to keep responses small and to avoid leaking auth material.
include_source_ref: bool = False — when true, source.source_ref is populated with a redacted reference: host/path for URLs (query string stripped, including signed tokens) or basename for local files. data_url and base64 inputs always return null for source_ref. Off by default to avoid leaking paths and signed URLs.

URL Handling

VISION_URL_MODE controls remote-image handling:

auto passes URLs through for analysis and comparison, but downloads them when inspection, cropping, or OCR requires image bytes.
passthrough prefers URL passthrough, except for tools that require bytes.
download always downloads and verifies remote images before model calls.

Downloads are streamed with byte limits, redirects are security checked, and downloaded or encoded inputs are verified as supported images.

Troubleshooting

If Claude Code cannot find the PyPI package:

UV_DEFAULT_INDEX=https://pypi.org/simple uvx --refresh vision-mcp

If the MCP server does not connect:

claude mcp get vision
uvx vision-mcp

If you change the Claude Code configuration:

claude mcp remove --scope user vision

Then add it again with the updated values.

Development

git clone https://github.com/666666999999666/vision_mcp.git
cd vision_mcp
python -m venv .venv
.venv/bin/pip install -e ".[dev]"
cp .env.example .env
.venv/bin/python -m unittest discover -s tests -v

License

MIT

Available Tools

6 tools

vision_analyzeA

Analyze an image using a vision-language model. Returns a unified JSON envelope wrapping summary, observations, uncertainties, and suggested follow-ups (see README 'Response format' for the full schema).

Supports URL, local file path, data URL, and Base64 input.

Task types guide the model:

general: General analysis (default)
qa: Answer a specific question about the image
ui: Analyze UI/layout/interactions/accessibility
chart: Analyze charts/graphs/data
document: OCR and document structure
object: Identify and locate objects
screenshot: Analyze application screenshots
code_screenshot: Read code from screenshots

ParametersJSON Schema

Name	Required	Default
`task`	No	general
`detail`	No	auto
`prompt`	No	请描述这张图片的内容。
`include_raw`	No
`image_source`	Yes
`include_source_ref`	No

Output Schema

ParametersJSON Schema

Name	Required	Description
`result`	Yes

TDQS

A3.8/5.0

Behavior3/5

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

With no annotations, the description must disclose behavioral traits. It mentions return format (JSON envelope with summary, observations, etc.) and input support, but does not explicitly state read-only nature, rate limits, or side effects. The reference to README helps, but behavioral transparency is adequate but not thorough.

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 concise and well-structured: a clear first sentence for purpose, followed by supported inputs and task types in a bullet-like list. No extraneous information, and all sentences add value.

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

Completeness3/5

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

The tool has 6 parameters and an output schema. The description covers purpose, inputs, and task types, and references the README for output schema. However, unexplained parameters (detail, prompt, include_raw, include_source_ref) and the absence of behavioral constraints make it somewhat incomplete for a complex tool.

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

Parameters2/5

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

Schema description coverage is 0%, so the description should compensate. It explains only 'image_source' and 'task' (task types), leaving 'detail', 'prompt', 'include_raw', and 'include_source_ref' unexplained. The default prompt in Chinese may confuse. Given the low coverage, the description adds insufficient parameter meaning.

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 the tool analyzes images using a vision-language model, enumerates supported input types, and lists specific task types (general, qa, ui, chart, document, object, screenshot, code_screenshot). This distinguishes it from sibling tools like vision_extract_text (OCR) and vision_compare, making the purpose specific and well-defined.

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 specifies task types that guide model behavior, implicitly advising when to use this tool (e.g., for general analysis, UI analysis, chart analysis). However, it does not explicitly state when not to use it or directly compare with alternatives, so guidance is present but not exhaustive.

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

vision_capabilitiesA

Return current vision-mcp server capabilities, supported models, and limits. Call this to discover what the server can do before using other tools.

ParametersJSON Schema

Name	Required	Description	Default
No parameters

Output Schema

ParametersJSON Schema

Name	Required	Description
`result`	Yes

TDQS

A4.5/5.0

Behavior3/5

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

No annotations are provided. The description suggests a read-only operation by using 'Return', but does not explicitly state it has no side effects or require authentication. More detail would help.

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?

Two concise sentences that front-load the purpose and usage. No unnecessary words.

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?

With no parameters and an output schema defined, the description sufficiently covers what the tool returns and when to use it. Complete for this simple discovery tool.

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?

No parameters exist; schema coverage is 100%. The description adds no parameter info because none are needed. Baseline score of 4 is appropriate.

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?

Clearly states it returns server capabilities, supported models, and limits. Distinguishes from sibling tools which are analysis tools, making this a discovery tool.

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

Usage Guidelines5/5

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

Explicitly advises to call this before using other tools, providing direct usage context.

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

vision_compareA

Compare two or more images and identify differences. Use for:

UI regression testing (before/after screenshots)
Design vs implementation comparison
Bug screenshot comparison
Version diff of documents

Returns structured differences with confidence levels.

ParametersJSON Schema

Name	Required	Default
`focus`	No	general
`prompt`	No	请比较这些图片的异同
`include_raw`	No
`image_sources`	Yes
`include_source_ref`	No

Output Schema

ParametersJSON Schema

Name	Required	Description
`result`	Yes

TDQS

A3.9/5.0

Behavior3/5

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

No annotations are provided, so the description carries full burden. It mentions returning structured differences with confidence levels, which adds transparency, but lacks details on side effects, prerequisites, or limitations (e.g., supported formats, performance).

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?

Description is concise with bullet points for use cases and ends with a clear statement about return values. Every sentence adds value, and the structure is front-loaded with the core purpose.

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

Completeness4/5

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

Given the tool's complexity (5 params, output schema exists), the description covers the main purpose and return format, but parameter guidance is lacking. Overall adequate for a comparison tool, though more detailed parameter descriptions would improve completeness.

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

Parameters2/5

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

Schema coverage is 0% (no parameter descriptions). The description does not explain parameters beyond the schema (names, types, defaults). For a tool with 5 parameters, this is insufficient; the description should at least clarify 'focus', 'prompt', 'include_raw', and 'include_source_ref'.

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?

Description clearly states the tool compares two or more images and identifies differences, listing specific use cases like UI regression testing and design comparison, which distinguishes it from sibling tools like vision_analyze or vision_extract_text.

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?

Description provides explicit use cases (UI regression, design vs implementation, bug screenshot, version diff) but lacks guidance on when not to use or alternatives; context from sibling tools implies other tools for analysis, but no exclusions are mentioned.

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

vision_crop_analyzeA

Crop a region of an image and analyze it with VLM. This is the most powerful tool for inspecting small text, UI elements, chart data, or error messages.

Coordinates are NORMALIZED (0.0 to 1.0), where (0,0) is top-left and (1,1) is bottom-right.

Workflow: Use vision_inspect first to get dimensions, then vision_analyze for overview, then vision_crop_analyze to zoom into specific regions of interest.

ParametersJSON Schema

Name	Required	Default
`x`	Yes
`y`	Yes
`task`	No	general
`width`	Yes
`height`	Yes
`prompt`	No	请详细描述这个区域的内容
`include_raw`	No
`image_source`	Yes
`include_source_ref`	No

Output Schema

ParametersJSON Schema

Name	Required	Description
`result`	Yes

TDQS

A4/5.0

Behavior3/5

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

No annotations are provided, so the description must disclose behavior. It explains normalized coordinates and the VLM analysis but does not state that the operation is non-destructive, mention authentication or rate limits, or describe potential latency. The ambiguity around 'crop' (virtual vs. actual modification) is not resolved.

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 composed of three short paragraphs, each with a clear focus: purpose, coordinate details, and workflow. It is concise without wasted words, though it could be slightly more structured with bullet points or sections.

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

Completeness3/5

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

Given 9 parameters and an output schema, the description covers the main behavior and coordinate system but omits explanation of several optional parameters and does not describe the return format. The workflow guidance adds context, but the lack of full parameter coverage affects completeness.

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 0%, so the description must compensate. It explains the critical coordinate parameters (x, y, width, height) are normalized 0-1, which is helpful. However, it does not explain other parameters like task, prompt, include_raw, or include_source_ref, leaving gaps for the agent.

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 the tool's action: 'Crop a region of an image and analyze it with VLM.' It also specifies use cases like inspecting small text, UI elements, and chart data, and distinguishes from sibling tools by positioning it as a zoom-in tool after broader analysis.

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

Usage Guidelines5/5

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

Explicit workflow guidance is provided: 'Use vision_inspect first to get dimensions, then vision_analyze for overview, then vision_crop_analyze to zoom into specific regions.' This tells the agent exactly when and in what order to use the tool.

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

vision_extract_textA

Extract visible text from an image using OCR. Returns structured text organized by reading order.

Use this for: screenshots with text, scanned documents, receipts, tables, forms, Chinese/English OCR, and any text-heavy images.

Uses a configured dedicated OCR model when enabled. If the dedicated OCR model is unavailable, automatically falls back to the VLM provider.

ParametersJSON Schema

Name	Required	Default
`language`	No	auto
`include_raw`	No
`image_source`	Yes
`preserve_layout`	No
`include_source_ref`	No

Output Schema

ParametersJSON Schema

Name	Required	Description
`result`	Yes

TDQS

A3.5/5.0

Behavior3/5

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

No annotations provided, so description carries full burden. Discloses use of dedicated OCR model with automatic fallback to VLM, and mentions output is structured by reading order. Lacks details on authentication, rate limits, or behavior with empty images.

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?

Three sentences, each with a distinct purpose: function definition, use cases, model behavior. No filler, front-loaded with the core action. Conciseness earned.

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?

Despite 5 parameters and 0% schema coverage, description omits parameter details and output specifics beyond 'structured text'. Output schema exists but not leveraged. Tool complexity warrants more explanation for correct invocation.

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

Parameters1/5

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

Schema description coverage is 0% and description does not explain any of the 5 parameters (language, include_raw, image_source, etc.). Adds no meaning beyond the schema itself, which also lacks descriptions. Baseline expectation: description should compensate.

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?

Clearly states it extracts text from images via OCR and returns structured text by reading order. Lists specific use cases (screenshots, documents, receipts, etc.) and distinguishes from sibling tools like vision_analyze which analyze images rather than extract text.

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?

Explicitly recommends using for text-heavy images with examples. Does not specify when not to use or provide alternative tools, but the context and sibling list imply guidance. Lacks exclusion criteria like 'if no text present'.

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

vision_inspectA

Inspect image metadata (dimensions, format, size, mode) without calling VLM. Use this before detailed analysis to understand the image dimensions and plan crop coordinates.

ParametersJSON Schema

Name	Required	Description	Default
`image_source`	Yes
`include_source_ref`	No

Output Schema

ParametersJSON Schema

Name	Required	Description
`result`	Yes

TDQS

A4.4/5.0

Behavior5/5

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

With no annotations, the description fully bears the burden. It explicitly states that the tool does NOT call a VLM, implying it is lightweight and non-destructive. This is critical behavioral context.

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?

Two sentences, no wasted words. The key information is front-loaded: purpose first, then usage guidance. Every sentence adds value.

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

Completeness4/5

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

For a simple metadata inspection tool with an output schema, the description covers purpose and usage well. However, the lack of parameter documentation is a minor gap. Overall, it is complete enough for the tool's simplicity.

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

Parameters2/5

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

Schema coverage is 0% (no descriptions in input schema), so description must compensate. However, the description does not explain what 'image_source' expects (e.g., file path, URL) or the effect of 'include_source_ref'. The agent receives no added meaning beyond parameter names.

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 it inspects image metadata (dimensions, format, size, mode) and distinguishes itself from detailed analysis tools like vision_analyze. The verb 'inspect' plus specific metadata fields make the purpose unambiguous.

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

Usage Guidelines5/5

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

Explicitly advises using this before detailed analysis to understand dimensions and plan crop coordinates. This provides clear context for when to use this tool versus siblings like vision_analyze or vision_crop_analyze.

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.

6 tool updatesv0.0.3Jul 15, 2026
- First observedvision_analyze
- First observedvision_capabilities
- First observedvision_compare
- First observedvision_crop_analyze
- First observedvision_extract_text
- First observedvision_inspect

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

A4.2/5.0

Scored 2026-07-15 12:45 across 6 tools

Disambiguation5/5

Each tool has a clearly distinct purpose: general analysis, comparison, region analysis, OCR, and metadata inspection. There is no overlap in functionality.

Naming Consistency5/5

All tool names follow the consistent pattern 'vision_<verb>' or 'vision_<verb>_<noun>', using lowercase and underscores. The naming is predictable and readable.

Tool Count5/5

With 5 tools, the server is well-scoped for image analysis. Each tool addresses a core need without unnecessary clutter.

Completeness5/5

The tool set covers all essential image analysis operations: general analysis, comparison, region analysis, OCR, and metadata inspection. The vision_capabilities task type also provides self-description.

Maintenance

ActivityMaintained

ResponsivenessSyncing

Resources

GitHub Repository

Need Help?

Related Servers

Related MCP Connectors

Frenchie A
io.github.Lab94
OCR, transcription, file extraction, and image generation for AI agents via MCP.
AI & Machine Learning Image & Video Processing Speech Processing
72
NinjaChat MCPOAuth
ai.ninjachat
Create images & video from any MCP agent — 17 models, spend limits, one URL.
Image & Video Processing AI & Machine Learning
1
Razi Media Tools A
io.github.razikallayi
Image processing over MCP: compress and split images, generate images and QR codes.
Image & Video Processing
41
Imejis AOAuth
io.imejis
Generate on-brand images from your AI agent: design, edit, and render templates over MCP.
App Automation Design Tools Image & Video Processing
2

Related MCP Servers

agent-vision-mcp
Image & Video Processing AI & Machine Learning
idealizing
A
license
Not graded
quality
D
maintenance
Provides image analysis, inspection, cropping, OCR, and comparison capabilities via the Model Context Protocol, allowing AI agents to process and manipulate images using vision models.
Updated 3 months ago (2026-06-16 10:49 UTC)
2
MIT
vision-bridge-mcp
Image & Video Processing AI & Machine Learning
G12789
A
license
A
quality
D
maintenance
Bridges a vision model to enable text-only models like DeepSeek to describe images, extract text, and compare images via MCP tools.
Updated 3 months ago (2026-06-15 09:15 UTC)
5
42 npm
9
MIT
Vision MCP Server
Image & Video Processing AI & Machine Learning
Loveacup
A
license
A
quality
D
maintenance
Enables AI agents to analyze images, extract text, compare images, and analyze video through any OpenAI-compatible vision model.
Updated 7 months ago (2026-02-09 08:40 UTC)
4
166 npm
20
MIT
VisionToolMCP
Image & Video Processing AI & Machine Learning
Winter-And-You-Gone
F
license
A
quality
B
maintenance
Enables text-only agents to process images by accepting image files, base64 data, or URLs, sending them to multimodal models, and returning structured text results via MCP.
Updated a month ago (2026-08-13 13:47 UTC)
4
1
-

vision-mcp

Features

Claude Code

Requirements

Add To Claude Code

Dedicated OCR Model

Other MCP Clients

Tools

Response format

Success envelope

Failure envelope

Per-tool result shape

Multi-image input

Opt-in flags

URL Handling

Troubleshooting

Development

License

Available Tools

Output Schema

Output Schema

Output Schema

Output Schema

Output Schema

Output Schema

Tool Schema Changelog

Maintenance

Resources

Related MCP Connectors

Related MCP Servers

Per-tool `result` shape