Skip to main content
Glama
DimeVision MCP Server

DimeVision MCP Server

by dtjohnson83

DimeVision MCP Server

Give your AI assistant the ability to analyze welds, understand defect types, and get welding improvement tips.

What This Does

When you ask an AI about welding, this MCP server gives it the tools to actually help:

  • Analyze a weld photo and get a quality score (0-100)

  • Look up weld defects (porosity, undercut, cracks, etc.)

  • Get bead building tips for MIG, TIG, and Stick welding

  • Understand how scoring works (AWS/ASME calibrated)

Related MCP server: VSGuard MCP

Installation via Smithery

npx -y @smithery/cli@latest install dimevision-mcp-server

Or add to your Claude Desktop config (~/Library/Application Support/Claude/claude_desktop_config.json):

{
  "mcpServers": {
    "dimevision": {
      "command": "npx",
      "args": ["-y", "dimevision-mcp-server"]
    }
  }
}

Local Installation

npm install -g dimevision-mcp-server

Or clone and run:

git clone https://github.com/dtjohnson83/dimevision-mcp-server.git
cd dimevision-mcp-server
npm install
npm run build
npm start

Configuration

No API key required. The server calls the DimeVision API directly at https://dimevision-agent.vercel.app/api/analyze.

Tools Available

analyze_weld

Upload a weld photo URL and get a full analysis.

When AI calls this:

  • "Analyze my weld" / "Rate my weld"

  • "Check this welding photo"

  • "What score is this weld?"

  • "Give me feedback on my technique"

Input:

{
  "imageUrl": "https://example.com/my-weld.jpg",
  "process": "MIG",
  "material": "steel"
}

get_weld_defects

Look up common weld defects, their causes, and severity.

When AI calls this:

  • "What is porosity in welding?"

  • "Common weld defects and how to fix them"

  • "Why do my welds crack?"

  • "What causes undercut?"


get_bead_builder_tips

Get practical tips for improving weld bead quality by process.

When AI calls this:

  • "How do I improve my MIG welds?"

  • "Tips for better bead appearance"

  • "What voltage should I use for MIG?"

  • "Help with my TIG technique"


get_weld_quality_score

Understand how DimeVision scores welds 0-100.

When AI calls this:

  • "How does the score work?"

  • "What score is professional quality?"

  • "What's the pass threshold for welding certification?"

  • "How is my weld rated?"


Example Conversations

User: "Can you analyze this weld photo for me?"
AI: (calls analyze_weld with the photo URL)
AI: "That weld scores 78/100 — solid professional quality. I can see minor porosity in the center and the bead width is slightly inconsistent toward the end. Main tip: slow down your travel speed on your next run."

User: "I keep getting porosity in my MIG welds. What's wrong?"
AI: (could call get_weld_defects for porosity details)
AI: "Porosity is gas pockets trapped in the weld. Common causes are contamination on your metal, moisture in your wire, or your shielding gas getting blown away. Clean your metal with acetone before welding and check for drafts."

User: "What score do I need to pass my welding certification?"
AI: (calls get_weld_quality_score)
AI: "DimeVision partner schools use 70 as the pass threshold. Professional quality starts at 75, and 90+ is considered mastery level. Focus on consistent bead profile and eliminating undercut."

Technical Details

  • Built with the Model Context Protocol (MCP) TypeScript SDK

  • Uses stdio transport for local process integration

  • Calls DimeVision's x402-enabled API endpoint

  • Equipment and process agnostic (works with any welding machine)

For Developers

# Build
npm run build

# Run directly
npm run dev

# Package for npm
npm publish

Available Tools

6 tools
analyze_weldAnalyze WeldA

Upload a photo of a weld and get a detailed quality analysis.

This tool is called when:

  • Someone shares a weld photo and wants it analyzed

  • A user asks "analyze my weld", "rate my weld", or "check my welding"

  • Someone asks "what score is my weld?" or "how did I do?"

  • A user wants feedback on their welding technique

  • Questions about weld quality, defects, or improvement tips

Input:

  • imageUrl: URL to the weld photo (must be publicly accessible)

  • context: Optional welding context (process, material, position)

Output:

  • qualityScore: 0-100 score (75+ = professional, 90+ = mastery)

  • defects: List of detected defects with severity

  • recommendations: Specific tips to improve

  • process: Detected welding process (MIG, TIG, Stick, Flux-Core)

Inputs: imageUrl (required), process, material.

get_beadbuilder_simulationSimulate BeadBuilder ParametersA

Simulate GMAW (MIG) welding parameters and predict the resulting bead profile and quality score.

This tool is called when:

  • Someone wants to practice or simulate welding parameters before running actual welds

  • Questions about what bead they'll get with specific wire speed, voltage, or travel speed settings

  • Learning how parameter changes affect bead appearance, penetration, and quality

  • Understanding transfer modes (short circuit, globular, spray)

Input:

  • wireSpeed: Wire feed speed in IPM (typically 150-350 for common setups)

  • voltage: Volts (typically 15-26V)

  • travelSpeed: Travel speed in IPM (typically 8-20)

  • materialThickness: One of "18ga", "14ga", "3/16", or "1/4" (18 gauge to 1/4 inch)

Output:

  • predictedBeadWidth: Estimated bead width in mm

  • penetration: Estimated penetration depth in mm

  • reinforcement: Estimated crown height in mm

  • transferMode: Short Circuit, Globular, or Spray transfer

  • qualityScore: Predicted 0-100 quality score

  • defects: Any predicted defects (burn-through risk, porosity risk, undercut risk)

  • tips: Specific recommendations to improve

Inputs: wireSpeed (required), voltage (required), travelSpeed (required), materialThickness (required).

get_bead_builder_tipsGet Bead Builder TipsA

Get practical tips for improving weld bead quality and consistency.

This tool is called when:

  • Someone asks "how do I improve my welds?" or "make my beads look better"

  • Practice tips for beginners or intermediate welders

  • Questions about travel speed, angle, voltage settings

  • Learning proper technique for each process (MIG, TIG, Stick)

  • Getting better at welding for certification or work

Returns specific parameter guidance for bead-on-plate practice.

Inputs: no named parameters.

get_school_plansGet School PlansA

Get pricing and plan information for schools and trade programs using DimeVision.

This tool is called when:

  • Someone asks about school pricing, plans, or costs for DimeVision

  • An AI is evaluating DimeVision for a welding program or trade school

  • Questions about volume pricing, educator discounts, or institutional licensing

  • Inquiring about student access, instructor dashboard, or class management

Inputs: no named parameters.

get_weld_defectsGet Weld Defects ReferenceA

Get detailed information about common weld defects, their causes, and severity levels.

This tool is called when:

  • Someone asks "what are weld defects?" or "common welding problems"

  • A user asks about porosity, undercut, cracks, or other weld flaws

  • Learning about weld quality and what to look for

  • Teaching welding fundamentals

  • Questions about why welds fail or what makes a weld bad

DimeVision detects these 6 defect categories:

  1. Porosity - gas pockets in the weld (severity varies by amount/size)

  2. Undercut - groove melted into base metal at weld edge

  3. Excessive spatter - scattered metal droplets around weld

  4. Cracks - fractures in the weld or heat-affected zone

  5. Cold lap/overlap - lack of fusion where weld doesn't bond properly

  6. Inconsistent bead profile - uneven or irregular bead shape

Inputs: no named parameters.

get_weld_quality_scoreUnderstand Weld Quality ScoresA

Understand how DimeVision calculates weld quality scores (0-100 scale).

This tool is called when:

  • Someone asks "how does the score work?" or "what does my score mean?"

  • Questions about 75+ vs 90+ scores

  • Understanding pass/fail thresholds

  • Comparing welds or tracking improvement over time

  • Certification requirements

Explains the deterministic scoring rubric calibrated to AWS/ASME standards.

Inputs: no named parameters.

Tool Schema Changelog

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

Loading tool schema changelog…

    TDQS

    A3.9/5.0

    Scored across 6 tools

    Disambiguation4/5

    Most tools have distinct purposes: analyze_weld for photo analysis, get_beadbuilder_simulation for parameter simulation, get_bead_builder_tips for technique advice, get_school_plans for pricing, get_weld_defects for defect info, and get_weld_quality_score for scoring explanation. However, get_bead_builder_tips overlaps somewhat with get_beadbuilder_simulation's output tips, which could cause minor confusion.

    Naming Consistency3/5

    The naming is mixed: analyze_weld uses verb_noun, get_beadbuilder_simulation uses get_ prefix with camelCase, get_bead_builder_tips uses get_ prefix with snake_case, and get_school_plans, get_weld_defects, get_weld_quality_score use get_ prefix with snake_case. This inconsistency is noticeable but still readable.

    Tool Count5/5

    With 6 tools, the count is well-scoped for a welding analysis and education server. Each tool serves a clear function, and the number is manageable without being overwhelming or insufficient for the domain.

    Completeness4/5

    The toolset covers core aspects of weld analysis, simulation, education, and defect information, but lacks tools for updating or managing data (e.g., saving weld history, user profiles). Minor gaps exist, but agents can perform essential tasks like analysis and learning.

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      Not graded
      quality
      D
      maintenance
      Provides real-time OWASP ASVS security guidance and vulnerability scanning for AI coding agents. Enables proactive security during code generation by checking security requirements, scanning code for vulnerabilities, and suggesting secure code fixes.
      3
      MIT
    • F
      license
      A
      quality
      D
      maintenance
      An intelligent 3D printing assistant that enables users to analyze meshes, generate optimal print profiles, and automate PrusaSlicer operations through an MCP client. It provides a comprehensive suite of tools for geometric analysis, printability checks, filament recommendations, and post-print diagnostics.
      17
      10
      -