collectible-vault
Enables AI assistants to manage a universal collectibles vault with natural language, supporting item lookup, addition, valuation, and portfolio summary via FastMCP.
README
Universal Collectibles Vault (collectible-vault)
An open-source, containerized, self-hosted web application for collectors to manage comics, Funko Pops, action figures, and trading cards. Replaces subscription services (CLZ, Funko tracking apps) with zero-friction camera scanning (UPC + local vision AI), automatic eBay sold-comps valuation, and FastMCP integration for natural-language AI assistants.
🏛️ System Architecture
┌─────────────────────────────────────────────────────────────────────────┐
│ User Mobile / Web UI │
│ (HTML5 Camera Scanner / Vision Upload / Dashboard) │
└────────────────────────────────────┬────────────────────────────────────┘
│ REST API / Base64 Image
┌────────────────────────────────────▼────────────────────────────────────┐
│ FastAPI App Core (Container 1) │
├────────────────────────────────────┬────────────────────────────────────┤
│ • Barcode/UPC Intake Engine │ • Valuation Refresh Engine │
│ • Pre-flight Metadata Form │ • Analytics & Dashboard Stats API │
└──────────────┬─────────────────────┴─────────────┬──────────────────────┘
│ SQLite Hybrid Schema │ Base64 Vision Payload
┌──────────────▼─────────────┐ ┌──────────────▼──────────────────────┐
│ SQLite Hybrid Vault DB │ │ Ollama Vision LLM (Container 3) │
│ (JSON Dynamic Metadata) │ │ (qwen2-vl / llava model endpoint) │
└──────────────▲─────────────┘ └─────────────────────────────────────┘
│ Direct SQLite access / ORM
┌──────────────┴─────────────┐
│ FastMCP Server Daemon │ <=========> Claude Desktop / Ollama Agent
│ (Natural Language AI) │ "Show top comic value gains"
└────────────────────────────┘
🚀 Key Features
-
Frictionless "Snap & Add" Intake:
- Path A (UPC Barcode): Decode box/cover barcodes with instant pre-populated metadata lookup.
- Path B (Vision AI LLM): Upload photos directly to a local vision LLM (Ollama
qwen2-vlorllava) to extract Title, Issue #, Publisher, Era, and estimated condition. - Pre-flight Confirmation UI: Review auto-populated metadata and AI confidence score before saving.
-
Universal Dynamic Hybrid Schema:
- Universal fields across all item types (
id,title,category,purchase_price,current_market_value,condition_grade,notes). - Dynamic
metadata_jsonfield for category-specific tags (issue number, variant, box number, card set, grading service).
- Universal fields across all item types (
-
Live Market Valuation Engine:
- Sold-comps algorithm simulating completed market listings to track true Fair Market Value (FMV).
- Keeps historical valuation trend points to plot financial gains/losses over time with interactive Chart.js line graphs.
-
FastMCP Natural Language AI Agent Server:
- Exposes vault tools (
get_vault_summary,add_item,query_item_market_value,list_top_collectibles,refresh_vault_valuations) to Claude Desktop or Ollama agents.
- Exposes vault tools (
🐳 Quickstart Deployment with Docker Compose
Prerequisites
- Docker & Docker Compose installed on Linux / Proxmox VM.
1. Clone & Launch Stack
git clone https://github.com/your-username/collectible-vault.git
cd collectible-vault
docker-compose up -d
2. Access Web Dashboard
Open your web browser and navigate to:
http://localhost:8000
Or use your Proxmox VM IP: http://<PROXMOX_SERVER_IP>:8000
3. Enable Vision LLM Model (Optional)
Pull the vision LLM into the Ollama container:
docker exec -it collectible-vault-ollama ollama pull qwen2-vl
🤖 FastMCP Server Setup for AI Assistants
Claude Desktop Integration (claude_desktop_config.json)
Add the following to your Claude Desktop config file:
{
"mcpServers": {
"collectible-vault": {
"command": "python",
"args": [
"/path/to/collectible-vault/mcp_server.py"
],
"env": {
"DATABASE_URL": "sqlite:////path/to/collectible-vault/vault.db"
}
}
}
}
Now ask Claude:
"What's the total value of my comic book vault?"
"Add a 1st Edition Charizard card bought for $400 with current market value $1,850."
🛠️ REST API Reference
| Method | Endpoint | Description |
|---|---|---|
GET |
/api/items |
List collectibles (filters: category, search, sort_by) |
POST |
/api/items |
Save new item to vault |
GET |
/api/items/{id} |
Get item detail & valuation timeline |
DELETE |
/api/items/{id} |
Delete item from vault |
POST |
/api/intake/vision |
Upload photo for Ollama Vision LLM parsing |
POST |
/api/intake/barcode |
Lookup UPC barcode metadata |
POST |
/api/valuation/refresh |
Trigger live eBay sold comps refresh |
GET |
/api/dashboard/stats |
Get aggregated portfolio stats & top items |
📄 License
Released under the MIT License.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。