What Is an MCP Server and Why Your Game Asset Tool Needs One

AI assistants are learning to use tools

Large language models like Claude and ChatGPT started as text generators. Then they learned to browse the web, run code, and read files. The next step is using the same software your team uses — project management tools, databases, design systems, and asset catalogs.

The Model Context Protocol (MCP) is the standard that makes this possible. Developed by Anthropic and adopted across the AI ecosystem, MCP defines how an AI assistant connects to an external tool, discovers what operations are available, and executes them with proper authentication. Think of it as a USB port for AI — a universal interface that lets any compatible assistant plug into any compatible tool.

How MCP works in practice

An MCP server exposes a set of tools — discrete operations that an AI assistant can call. Each tool has a name, a description, and a schema defining its inputs and outputs. When an AI assistant connects to an MCP server, it:

1. Discovers available tools — reads the list of operations and their descriptions
2. Understands context — uses tool descriptions to decide which operation fits the user's request
3. Executes operations — calls tools with structured parameters and receives structured responses
4. Composes workflows — chains multiple tool calls to accomplish complex tasks

The assistant handles natural language understanding, intent mapping, and response formatting. The MCP server handles data access, validation, and business logic. Neither needs to know the internals of the other.

What PolyDrobe's MCP server provides

PolyDrobe exposes an MCP server at https://polydrobe.app/mcp with 11 tools covering the core asset management operations:

Reading data

• list_projects — returns all projects the authenticated user has access to
• get_project — returns project details including taxonomy (categories, rarities, statuses, tags)
• list_assets — returns assets in a project with filtering by category, rarity, status, and search query
• get_asset — returns full asset details including all variants
• list_variants — returns variants for a specific asset
• search — full-text search across assets and variants in a project

Creating data

• create_asset — creates a new asset with name, category, rarity, description, and tags
• create_variant — creates a new variant with name, price, status, rarity, and metadata

Media access

• get_variant_thumbnail — returns the thumbnail URL for a variant
• get_variant_image — returns a specific image from a variant's gallery
• set_variant_thumbnail — sets or updates a variant's thumbnail

Every operation is project-scoped and respects the same role-based access control as the web interface. An AI assistant connected with an Editor's token can create assets; one connected with a Viewer's token can only read.

What you can actually do with this

The real value is not the individual tools — it is what happens when an AI assistant can compose them into workflows driven by natural language.

Ask questions about your catalog

"How many variants in Project Orion are still in Review status?"

The assistant calls list_projects to find Project Orion, then list_assets with a status filter, aggregates the variant counts, and returns a number with context.

Create assets from descriptions

"Create an asset called Night Ops Vest in the Outerwear category with Epic rarity. Add three variants: Midnight Black, Forest Green, and Arctic White, all priced at 1500 coins."

The assistant calls create_asset once, then create_variant three times with the right parameters. What would take a minute of clicking through forms happens in one sentence.

Search and compare

"Find all Legendary-rarity helmets and list their prices."

The assistant calls search or list_assets with category and rarity filters, extracts pricing from each variant, and presents a comparison table.

Generate reports

"Give me a release readiness summary for Release 2.1 — how many variants are Approved vs still In Progress?"

The assistant queries the project, filters by release version, groups by status, and produces a summary that would otherwise require manual filtering and counting.

How to connect

PolyDrobe's MCP server uses OAuth 2.1 for authentication. To connect an AI assistant:

1. Get a Bearer token — authenticate via PolyDrobe's OAuth endpoint at /o/token/
2. Point the assistant to the MCP endpoint — https://polydrobe.app/mcp
3. Start asking questions — the assistant discovers available tools automatically

For Claude, add the MCP server in your Claude configuration. For other assistants that support MCP or OpenAPI, the plugin manifest at /.well-known/ai-plugin.json and the OpenAPI spec at /api/openapi.json provide the connection details.

Why this matters for game teams

Game asset management involves a lot of repetitive queries and data entry. "What is the status of the Season 3 battle pass items?" should not require logging in, navigating to the right project, setting filters, and scanning a list. It should be a question you ask your AI assistant while reviewing a design doc or sitting in a planning meeting.

MCP also opens the door to automation. A CI pipeline that uses an AI agent to verify asset completeness before a release build. A Slack bot that answers "what shipped in the last update?" by querying the catalog. A content planning tool that pulls the current catalog state to identify gaps.

PolyDrobe is one of the first game asset management tools to ship an MCP server. As AI assistants become standard development tools, the catalog that can talk to them has a structural advantage over the one that cannot.

Key takeaways

• MCP (Model Context Protocol) is a standard that lets AI assistants connect to external tools and execute operations.
• PolyDrobe's MCP server exposes 11 tools for reading, creating, and searching game assets and variants.
• AI assistants can compose these tools into workflows: catalog queries, batch asset creation, release readiness reports.
• Authentication uses OAuth 2.1 with the same role-based access control as the web interface.
• The MCP server is live at https://polydrobe.app/mcp — connect any MCP-compatible AI assistant to start using it.