Skip to main content
Glama

Vision QA MCP

An MCP server that gives an AI agent automated quality control for generated images. After a model produces an image, the agent calls qa_check and gets back a structured verdict — character accuracy, style consistency, quality, and composition, scored against your reference images and rules — powered by Claude vision.

Built for AI media pipelines where a human can't eyeball every frame: generate → QA → regenerate-if-failed, automatically.

Why

Image models drift. A character loses a feature, the style wobbles, a hand comes out wrong. In a production pipeline that ships dozens of images, you need the agent itself to catch this before a human ever sees it. This server makes "QA every image" a single tool call with a pass/fail and actionable issues — and it pairs naturally with claude-vision-mcp (let the agent see) for a full see-and-verify loop.

Related MCP server: image-forensics-mcp

Tools

  • qa_check — review one image and return a structured verdict:

    • passed (bool), overall_score, and per-axis scores (character_accuracy, style_consistency, quality_score, composition_score) on a 0–1 scale

    • issues — a list of {severity, category, description, recommendation} (severity: critical / warning / minor)

    • should_regenerate (bool) and a one-line notes summary

    • Fails automatically on any critical issue, regardless of score.

  • list_scene_types — the supported scene_type values and what each expects.

You pass your own rules and references, so it works for any project:

qa_check(
  image_path="/path/to/generated.png",
  reference_images=["/refs/hero_front.png", "/refs/hero_face.png"],
  character_rules="The pilot has NO eyebrows in this form. Jacket has horizontal stripes, not clouds.",
  style_notes="High-contrast anime cel shading, cosmic purple lighting.",
  scene_type="solo",          # solo | portrait | battle | combat | group | action | interview
  pass_threshold=0.7,
)

scene_type adjusts composition expectations — e.g. a portrait may face the camera, while battle characters should face each other.

Requirements

  • Python ≥ 3.10

  • An Anthropic API key (ANTHROPIC_API_KEY)

  • Dependencies: mcp[cli], anthropic, Pillow

Install

git clone https://github.com/wonderstone843/vision-qa-mcp.git
cd vision-qa-mcp
pip install -e .
export ANTHROPIC_API_KEY=sk-ant-...

Use with Claude Code

claude mcp add vision-qa -- vision-qa-mcp

Or add to your MCP config:

{
  "mcpServers": {
    "vision-qa": { "command": "vision-qa-mcp" }
  }
}

Then instruct your agent: "After generating each image, run qa_check against the character refs; regenerate any that don't pass."

Configuration

  • ANTHROPIC_API_KEY (required)

  • ANTHROPIC_MODEL (optional, default claude-opus-4-8) — for QA on every generation, claude-haiku-4-5 or claude-sonnet-4-6 are cheaper and usually sufficient.

How it works

The image is downscaled to stay under the vision API limits, sent to Claude alongside any reference images and your rules, and the model is forced to call a submit_qa tool whose schema defines the four scores plus the issues list — so the output is always structured and parseable. The pass decision is overall_score >= pass_threshold AND no critical issues.

vision_qa_mcp/
  server.py   FastMCP server: qa_check + list_scene_types
  review.py   prompt, scoring rubric, forced-tool call to Claude vision
  images.py   downscale + base64-encode for the vision API

License

MIT — see LICENSE. Author: Joshua Penn.

Available Tools

2 tools
list_scene_typesA

List the supported scene_type values and what each one expects.

Inputs: no named parameters.

qa_checkA

Run a quality-control check on a generated image and return a structured verdict.

Call this immediately after generating an image, before showing it to the user.

Parameters:

  • image_path: Absolute path to the generated image to review.

  • reference_images: Optional list of reference image paths to compare against (e.g. character/style references). Up to 3 are used.

  • character_rules: Free-text rules the image must follow, e.g. "The pilot has NO eyebrows; the jacket has horizontal stripes."

  • style_notes: Free-text description of the expected visual style.

  • scene_type: One of solo, portrait, battle, combat, group, action, interview. Adjusts composition expectations (e.g. portraits may face camera).

  • pass_threshold: Minimum overall score (0-1) to pass. Default 0.7.

Returns a dict with: passed (bool), overall_score, character_accuracy, style_consistency, quality_score, composition_score (all 0-1), issues (list of {severity, category, description, recommendation}), should_regenerate (bool), notes, and model.

Inputs: image_path (required), scene_type, style_notes, pass_threshold, +2 more.

Tool Schema Changelog

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

Loading tool schema changelog…

    TDQS

    A4.1/5.0

    Scored across 2 tools

    Disambiguation5/5

    The two tools have completely distinct purposes: one lists scene types and their expectations, the other performs a quality check on images. There is no overlap or ambiguity.

    Naming Consistency4/5

    Both tools use snake_case, but one follows a verb_noun pattern (list_scene_types) while the other uses an abbreviated prefix (qa_check). The pattern is mostly consistent but has a minor deviation.

    Tool Count3/5

    With only 2 tools, the server is very focused. While it covers its core purpose, the count is on the low end and may feel thin for some use cases.

    Completeness3/5

    The tool set provides the essential functions: listing scene types and running a QA check. However, there is no tool to inspect or manage scene types beyond listing, and no tool to modify thresholds or references, leaving minor gaps.

    Maintenance

    ActivityInactive
    ResponsivenessNo recent community issues

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      A
      quality
      D
      maintenance
      An MCP server that provides classic computer vision feature extraction tools (color, texture, shape) for image analysis, enabling LLMs to perform precise mathematical image comparisons and quality assessments.
      9
      MIT
    • A
      license
      Not graded
      quality
      D
      maintenance
      An MCP server for AI image generation and editing, allowing users to generate images from text, edit existing images with instructions, and manage iterative editing sessions with undo/redo.
      76 npm
      1
      MIT
    • A
      license
      Not graded
      quality
      D
      maintenance
      An MCP server that fixes, validates, and generates visual text content for AI coding assistants.
      MIT