The Culture MCP Server
Provides tools for calling Hugging Face models for fashion-related tasks such as style archetype classification, content moderation, trend forecasting, ad CTR prediction, and recommendation.
Provides tools for interacting with Supabase database operations including user context, post context, trending posts, messaging, product tags, and community context.
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
The Culture MCP Server
MCP server for The Culture — a fashion-focused social media app. Exposes Supabase database operations and Hugging Face model inference as MCP tools for use in Claude Code, n8n agents, and the Larry orchestrator.
Tools
DB Tools (Supabase)
Tool | Description |
| Profile + last 20 saves + last 20 likes with product_tags |
| Single post content, image_url, product_tags |
| Posts filtered by archetype, ordered by recency |
| Insert a DM into messages (used by Larry) |
| Update a post's product_tags array |
| Community details, member count, recent posts |
Model Tools (Hugging Face)
Tool | Model | Task |
|
| Image → style archetype |
|
| Text → safe/unsafe + sub_category |
|
| Tabular → trend lifecycle stage |
|
| Tabular → click probability |
|
| User + posts → ranked affinity scores |
Related MCP server: mcp-n8n-builder
Running the server
This server uses the MCP Streamable HTTP transport (stateless mode). It listens on PORT (default 3000) and exposes:
POST /mcp— MCP JSON-RPC endpointGET /health— health check, returns{"status":"ok"}
npm install
npm run build
SUPABASE_URL=... SUPABASE_SERVICE_ROLE_KEY=... HF_TOKEN=... HF_USERNAME=Dc-4nderson npm startFor local development without a build step: npm run dev (uses tsx).
Deploying
Deploy anywhere that runs a Node HTTP server (Render, Railway, Fly.io, etc.). Set the same four env vars (SUPABASE_URL, SUPABASE_SERVICE_ROLE_KEY, HF_TOKEN, HF_USERNAME) in the platform's environment config, and make sure the platform's assigned PORT is respected (it is, via process.env.PORT).
Connecting a client (Claude Code / Claude Desktop)
Point the client at the deployed /mcp URL:
{
"mcpServers": {
"the-culture": {
"url": "https://your-deployment.example.com/mcp"
}
}
}Notes
Models not yet deployed to HF Inference Endpoints will return a 503 until deployed from their training notebooks.
CTR model default threshold is 0.3 (not 0.5) due to the platform's ~2.5% base click rate.
Recommendation engine requires user_id and post_ids that exist in the model's training data.
Related MCP Connectors
MCP server for interacting with the Supabase platform
MCP server for your apps' tools and custom tools, plus hosted AI agents and approval-gated workflows
MCP server for AI agents to plan, verify, and deploy Cloudflare-native apps.
Hosted MCP server for AI agents to post to LinkedIn profiles and X and upload to YouTube.
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceA Model Context Protocol server that enables Claude and other LLMs to perform database operations and invoke Edge Functions within Supabase through natural language.980 npm4MIT
- AlicenseAqualityDmaintenance🪄 MCP server for programmatic creation and management of n8n workflows. Enables AI assistants to build, modify, and manage workflows without direct user intervention through a comprehensive set of tools and resources for interacting with n8n's REST API.1012 npm86MIT
- FlicenseNot gradedqualityNot gradedmaintenanceA production-grade MCP server that centralizes AI integrations for Appwrite, Supabase, and MySQL databases. It also enables triggering n8n workflows via webhooks and performing generic REST API calls with built-in retry logic.-
- AlicenseNot gradedqualityDmaintenanceA production-ready MCP server providing AI assistants with intelligent Supabase database access, featuring dynamic schema discovery, complete user management, and file storage operations.1MIT