Skip to main content
Glama
Able-rip

agent-mosaic-skill

by Able-rip

BMBrick Photo-to-LEGO-Mosaic Skill

npm version License

Convert any local photo into a LEGO-style brick mosaic preview — directly from your AI agent.

@bmbrick/photo-to-lego-mosaic-skill is an MCP server that lets Claude, Cursor, Windsurf, and other AI agents transform local images into high-fidelity 3D brick mosaics using the same canonical engine as bmbrick.com.

Renamed from @bmbrick/agent-mosaic-skill (the old name still works but is deprecated — please migrate to the new name for future updates).

Visual Previews

Generated with the generate_lego_mosaic tool (3D render mode):

Square Tiles (square_1x1)

Original Photo

48x48 Preview

64x64 Preview

Original

48x48

64x64

Round Tiles (round_1x1)

Original Photo

48x48 Preview

64x64 Preview

Original

48x48 Round

64x64 Round

For more case studies and mosaic examples, check out our Pinterest Showcase.

Related MCP server: Brick Builder MCP App

Quick Start

npx -y @bmbrick/photo-to-lego-mosaic-skill

MCP Client Configuration

Claude Desktop (claude_desktop_config.json):

{
  "mcpServers": {
    "lego-mosaic": {
      "command": "npx",
      "args": ["-y", "@bmbrick/photo-to-lego-mosaic-skill"]
    }
  }
}

Cursor (.cursor/mcp.json):

{
  "mcpServers": {
    "lego-mosaic": {
      "command": "npx",
      "args": ["-y", "@bmbrick/photo-to-lego-mosaic-skill"]
    }
  }
}

Tool: generate_lego_mosaic

Parameter

Type

Required

Default

Description

imagePath

string

Yes

Absolute path to a local image

columns

number

No

48

Preview width in studs (max 64)

rows

number

No

48

Preview height in studs (max 64)

materialMode

string

No

square_1x1

square_1x1 or round_1x1

renderMode

string

No

3D

Render style

removeBackground

boolean

No

false

Clean up white/transparent backgrounds

Returns: a local watermarked PNG preview + a link to unlock the full project at bmbrick.com.

Unlock Full Project

The preview is watermarked and intended for creative exploration. To get the complete deliverables — HD unwatermarked mosaic, brick-by-brick PDF instructions, parts list with ordering links, and CSV/XML placement matrices — visit bmbrick.com.

Features

  • Same Engine as bmbrick.com — canonical color science, quantization, and 3D rendering pipeline

  • Local & Private — all processing happens on your machine via Node.js

  • Agent-Ready — ships with a SKILL.md for seamless agent workflow integration

  • 3D Render Mode — InstancedMesh PBR rendering with realistic brick studs

Engineering Docs

Deep technical reference for the color quantization pipeline that powers this skill and bmbrick.com:

  • Color quantization pipeline: principles, history, and lessons — current architecture, architectural decisions, verified principles, failed experiments from 25+ versions of iteration, and parameter sensitivity reference.

Migrating from @bmbrick/agent-mosaic-skill

If you were using the old name, the migration is one search/replace:

  1. Package name: @bmbrick/agent-mosaic-skill@bmbrick/photo-to-lego-mosaic-skill

  2. MCP tool name (in your agent's tool list): generate_bmbrick_mosaicgenerate_lego_mosaic

  3. Binary name (rare): agent-mosaic-skillphoto-to-lego-mosaic-skill

The tool parameters (imagePath, columns, rows, etc.) and return shape are unchanged.

Development

License

Dual-licensed:

  • MIT — MCP wrapper, skill/ directory, agent integration surface

  • Restricteddist/ engine files: free to use and distribute as-is, but no de-obfuscation, reverse engineering, or redistribution of modified versions

See LICENSE for full terms.

Related MCP Connectors

Related MCP Servers

  • A
    license
    A
    quality
    D
    maintenance
    Identifies LEGO parts, sets, and minifigures from local image files using the Brickognize API. It provides specialized tools for specific item recognition and integrates LEGO identification capabilities into MCP-enabled environments.
    5
    8 npm
    10
    MIT
  • A
    license
    Not graded
    quality
    C
    maintenance
    A Three.js MCP server for designing 3D brick constructions, enabling AI or interactive placement of bricks through natural language and providing tools to manage and export scenes.
    23
    MIT