Lumnix
AI-powered e-commerce research tools via MCP. Search Amazon, Alibaba & AliExpress for winning products, vet suppliers, and calculate FBA margins. 19 tools, free tier available.
README
<div align="center">
🔍 Lumnix
AI-powered e-commerce research tools via MCP
Stop paying for three different SaaS dashboards. Ask your AI to find winning products, vet suppliers, and calculate margins — all in one conversation.
npm install lumnix
Get API Key · Documentation · npm Package
</div>
Why Lumnix
The old way: Subscribe to Jungle Scout for product research ($49/mo), Keepa for price history ($50/mo), some Chrome extension for Alibaba vetting ($20/mo), and a spreadsheet for margin math. Switch between tabs all day. Miss opportunities.
The Lumnix way: One API key. 19 tools. In your AI assistant. No dashboards. No Chrome extensions. No tab-hopping.
"Find me wireless earbuds under $20 with high sales volume, low reviews, and a supplier on Alibaba with MOQ under 500"
Your AI calls Lumnix. You get structured data with revenue estimates, competition analysis, and profit margins. You decide. Done.
What makes it different
- 🚫 No dashboards — Works inside Claude, Cursor, Windsurf, Cline, or any MCP client
- 📊 Keepa-powered — Real price history and BSR tracking, not estimates
- 🎯 Opportunity Score — 0-100 niche score based on demand, competition, and margin potential
- 🏭 Full sourcing pipeline — Amazon → Alibaba → AliExpress, all connected
- 💰 Profit calculator — Full P&L with FBA fees, shipping, PPC ads, refunds, and ROI
- ⚡ 30-second setup — No install. No config file editing. Just an API key and one JSON block
⚡ Quick Start
1. Get an API key (free)
Sign up at lumnix.dev — no credit card required. Get your key in 30 seconds.
2. Add to your MCP client
Claude Desktop — edit ~/Library/Application Support/Claude/claude_desktop_config.json:
{
"mcpServers": {
"lumnix": {
"command": "npx",
"args": [
"mcp-remote",
"https://lumnix.dev/api/mcp",
"--header",
"Authorization: Bearer YOUR_API_KEY"
]
}
}
}
⚠️ Claude Desktop doesn't support remote MCP URLs natively. The
mcp-remotebridge (auto-installed by npx) makes it work. Requires Node.js installed.
Cursor — Settings → MCP → Add new global MCP server:
{
"mcpServers": {
"lumnix": {
"url": "https://lumnix.dev/api/mcp",
"headers": {
"Authorization": "Bearer YOUR_API_KEY"
}
}
}
}
Windsurf / Cline / any MCP client — same pattern as Cursor. URL + Bearer token header.
3. Start researching
Restart your AI client. Then just talk naturally:
"Search Amazon for portable blenders with high sales and low reviews"
That's it. No Docker. No config files to debug.
🛠 Tools Reference
Lumnix gives your AI 19 specialized tools across three platforms:
Amazon — 11 tools
| Tool | What it does |
|---|---|
amazon_search_products |
Search by keyword with price, rating, and review filters |
amazon_get_product_details |
Full product page — features, specs, reviews, images, category |
amazon_estimate_sales |
Monthly sales volume with confidence range |
amazon_opportunity_score |
⭐ 0-100 niche score — demand, competition, margin potential |
amazon_analyze_competition |
Top players, review counts, price ranges for a keyword |
amazon_best_sellers |
Best-selling products by category with sales data |
amazon_keyword_suggestions |
Amazon autocomplete keywords with search frequency |
amazon_price_history |
📊 Keepa-powered — historical price chart |
amazon_bsr_history |
📊 Keepa-powered — Best Sellers Rank trend over time |
amazon_analyze_negative_reviews |
Complaint categories, sentiment breakdown, common issues |
amazon_analyze_listing_quality |
CTR + CVR scoring with specific improvement suggestions |
Alibaba — 4 tools
| Tool | What it does |
|---|---|
alibaba_search_products |
Search suppliers and products with MOQ and price filters |
alibaba_get_product_details |
Pricing tiers, specs, variants, supplier credentials |
alibaba_vet_supplier |
Trust score, risk flags, verification status — avoid bad suppliers |
alibaba_match_to_amazon |
Match Alibaba products to Amazon listings for FBA arbitrage |
AliExpress — 4 tools
| Tool | What it does |
|---|---|
aliexpress_search_products |
Search with price, rating, and store filters |
aliexpress_get_product_details |
Variants, specs, shipping options, store ratings |
aliexpress_analyze_supplier |
Store score, feedback rate, top-rated status |
aliexpress_profit_calculator |
💰 Full P&L — product cost, shipping, FBA fees, PPC, refunds, ROI |
💬 Example Prompts
Product Research
"Search Amazon for stainless steel water bottles between $15-30 with at least 4 stars and under 1000 reviews"
Opportunity Scoring
"Score the 'portable blender' niche on Amazon. Is it worth entering?"
Competition Analysis
"Analyze the competition for 'organic dog treats'. How many sellers have less than 500 reviews?"
Keepa Price History
"Show me the price history for ASIN B0FHWKYT89. Has it been trending down?"
Supplier Vetting
"Find aluminum water bottle suppliers on Alibaba with trust score above 70 and MOQ under 500"
Margin Calculation
"I want to sell this AliExpress product on Amazon for $29.99. Calculate my profit margin including FBA fees, PPC ads at 15%, and 5% refund rate"
Full Workflow
"Find a product on Amazon with high sales volume and low reviews, find the matching supplier on Alibaba, vet them, and tell me if the margins work"
📊 Example Response
{
"asin": "B0FHWKYT89",
"title": "Portable Blender for Shakes and Smoothies",
"price": "$29.99",
"rating": 4.5,
"reviews": 2847,
"sales_volume": "5K+ bought in past month",
"monthly_revenue_est": "$149,950",
"is_amazon_choice": true,
"is_prime": true,
"number_of_offers": 12
}
💰 Plans
| Free | Pro | Business | |
|---|---|---|---|
| Price | $0 | $29/mo | $99/mo |
| API calls | 10/month | 5,000/month | 25,000/month |
| Platforms | Amazon only | Amazon + Alibaba + AliExpress | Amazon + Alibaba + AliExpress |
| Tools | Search, details, keywords | All 19 including Keepa | All 19 including Keepa |
| API keys | 2 | 5 | 20 |
| Keepa price history | ❌ | ✅ | ✅ |
| Opportunity score | ❌ | ✅ | ✅ |
| Supplier vetting | ❌ | ✅ | ✅ |
| Profit calculator | ❌ | ✅ | ✅ |
The free tier gives you 3 Amazon search tools — enough to see the data quality. When you're ready for Keepa-powered insights, opportunity scoring, and full sourcing across all platforms, upgrade here.
🔧 API Reference
Lumnix is a remote MCP server. No installation, no Docker, no local dependencies.
Endpoint: https://lumnix.dev/api/mcp
Auth: Bearer token in the Authorization header
Protocol: JSON-RPC 2.0 (MCP spec 2025-03-26)
Direct API call
curl -X POST https://lumnix.dev/api/mcp \
-H "Content-Type: application/json" \
-H "Authorization: Bearer YOUR_API_KEY" \
-d '{
"jsonrpc": "2.0",
"id": 1,
"method": "tools/call",
"params": {
"name": "amazon_search_products",
"arguments": { "keyword": "wireless earbuds", "max_price": 50 }
}
}'
🌐 Find Us On
Lumnix is listed on all major MCP directories:
- Smithery.ai — Search "lumnix"
- MCP Servers — Search "lumnix"
- Claude MCP Directory — Coming soon
- npm —
npm install lumnix·npm info lumnix
❓ FAQ
What MCP clients work with Lumnix? Any client that supports remote MCP servers with HTTP transport: Claude Desktop, Cursor, Windsurf, Cline, VS Code (with MCP extension), and more.
Is the free tier really free? Yes. 10 API calls/month, no credit card. Enough to search products and see the data quality before committing.
Where does the data come from? Amazon data via web APIs + Keepa (for price/BSR history). Alibaba and AliExpress via their respective platforms. All structured, enriched, and ready to act on.
How fast are the responses? Most tools respond in 1-5 seconds. Deep analysis tools (competition, reviews, opportunity score) can take up to 15 seconds.
Can I use Lumnix in my own app? Yes. The REST API works with any HTTP client. Call the MCP endpoint directly with your API key — no AI assistant required.
Do I need to install anything? No. Lumnix is a remote server. Just add the config to your MCP client and go.
🔒 Security
- API keys hashed with SHA-256 — never stored in plaintext
- Row-level security on all database tables
- Per-platform rate limiting
- Webhook signature verification for all payment events
📄 License
MIT
<div align="center">
Built by Molt Studios
</div>
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。