cryptosense-mcp
Real-time crypto market intelligence MCP. Get prices, trending coins, market overview, top coins by market cap, and portfolio value — all through natural language in Claude, Cursor, or any MCP client
README
CryptoSense MCP
Real-time crypto market intelligence for AI assistants.
CryptoSense MCP wraps the CoinGecko free API (no key needed) into a production-ready Model Context Protocol server built with FastMCP. Connect it to Claude, Cursor, Windsurf, or any MCP-compatible client and ask natural-language questions about crypto markets.
What this MCP does
| Tool | Description |
|---|---|
price |
Current price, market cap, volume & 24h change for any coin |
trending |
Top 10 trending coins by search volume (last 24 h) |
market_overview |
Global market cap, BTC/ETH dominance, 24h change |
top_coins |
Top N coins by market cap with full stats |
compare |
Side-by-side comparison of 2+ coins |
portfolio_value |
USD value of your holdings with best/worst performer |
All tools require a CryptoSense API key (see Authentication).
Installation
Option A — local with uv (recommended)
# 1. Clone
git clone https://github.com/your-org/cryptosense-mcp.git
cd cryptosense-mcp
# 2. Create venv and install
uv venv && uv pip install -e .
# 3. Copy and edit environment variables
cp .env.example .env
# Edit .env: set CMC_API_KEY if you have one, adjust MCP_PORT if needed
# 4. Generate your first API key
python -c "
import asyncio
from src.cryptosense.auth import generate_api_key
key = asyncio.run(generate_api_key('you@example.com'))
print('Your API key:', key)
"
# 5. Start the server
cryptosense-mcp
# or: python -m cryptosense.server
Option B — local with pip
pip install -e .
cp .env.example .env
python -m cryptosense.server
Option C — Docker
docker build -t cryptosense-mcp .
docker run -p 8000:8000 \
-e CMC_API_KEY=your_key \
-v cryptosense-data:/app/data \
cryptosense-mcp
Authentication
Every tool call requires an api_key parameter with a valid CryptoSense key.
Generate a key
import asyncio
from cryptosense.auth import generate_api_key
key = asyncio.run(generate_api_key(email="you@example.com", plan="free"))
print(key) # cs_Abc123...
Keys are stored in keys.db (SQLite). The keys.db file lives next to the server process (or at DATABASE_URL from .env).
CoinGecko API key (optional)
CoinGecko's free public API works without a key. If you experience rate limiting (30 calls/min on the free tier), sign up at https://www.coingecko.com/en/api for a free Demo API key and add it to your .env:
CG_API_KEY=CG-xxxxxxxxxxxxxxxxxxxx
The server currently uses the public endpoint. If you add a key, pass it via the
x-cg-demo-api-keyheader in_fetch()calls.
Configure Claude Desktop
Add to ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):
{
"mcpServers": {
"cryptosense": {
"command": "python",
"args": ["-m", "cryptosense.server"],
"cwd": "/absolute/path/to/cryptosense-mcp",
"env": {
"MCP_HOST": "127.0.0.1",
"MCP_PORT": "8000"
}
}
}
}
Or if the server is already running remotely, use the HTTP transport URL:
{
"mcpServers": {
"cryptosense": {
"url": "http://localhost:8000/mcp"
}
}
}
Configure Cursor
Open Settings → MCP → Add new MCP server and enter:
| Field | Value |
|---|---|
| Name | CryptoSense |
| Type | HTTP |
| URL | http://localhost:8000/mcp |
Or add to ~/.cursor/mcp.json:
{
"mcpServers": {
"cryptosense": {
"url": "http://localhost:8000/mcp"
}
}
}
Configure Windsurf
Add to ~/.codeium/windsurf/mcp_config.json:
{
"mcpServers": {
"cryptosense": {
"serverUrl": "http://localhost:8000/mcp"
}
}
}
Tool Reference & Example Prompts
price — Get coin price
"What is the price of Bitcoin?" "How much is Ethereum worth in EUR?" "Show me Solana's 24h change and market cap."
price(coin="bitcoin", currency="usd", api_key="cs_...")
# → { "coin": "bitcoin", "price": 67420.0, "market_cap": 1.32T, "change_24h_percent": 2.4, ... }
top_coins — Top coins by market cap
"Show me top 10 coins." "What are the top 20 cryptocurrencies by market cap?" "List the 5 biggest coins in EUR."
top_coins(limit=10, currency="usd", api_key="cs_...")
# → { "coins": [{ "rank": 1, "name": "Bitcoin", "price": 67420, ... }, ...] }
trending — Trending now
"What is trending in crypto today?" "Which coin is everyone searching for?" "Show me the hottest altcoins right now."
trending(api_key="cs_...")
# → { "trending_coins": [{ "name": "Pepe", "symbol": "PEPE", "market_cap_rank": 54, ... }] }
portfolio_value — Portfolio calculator
"Calculate my portfolio: 0.5 BTC, 5 ETH, 100 SOL." "How much is my crypto worth? I have 1 bitcoin and 10 ethereum." "What is my total if I hold 0.1 BTC, 500 DOGE, and 2 ETH?"
portfolio_value(
holdings={"bitcoin": 0.5, "ethereum": 5, "solana": 100},
currency="usd",
api_key="cs_...",
)
# → { "total_value": 54230.00, "best_performer": {...}, "breakdown": [...] }
compare — Side-by-side comparison
"Compare Bitcoin and Ethereum." "Which performs better: Solana, Avalanche, or Polkadot?" "Show me BTC vs ETH vs BNB."
compare(coins=["bitcoin", "ethereum", "solana"], currency="usd", api_key="cs_...")
# → { "comparison": [...], "best_performer_24h": "solana", "worst_performer_24h": "bitcoin" }
market_overview — Global snapshot
"What is the total crypto market cap?" "What is Bitcoin's market dominance today?" "Give me a global crypto summary."
market_overview(api_key="cs_...")
# → { "total_market_cap_usd": 2.45T, "btc_dominance_percent": 52.3, ... }
Environment Variables
| Variable | Default | Description |
|---|---|---|
CMC_API_KEY |
— | CoinMarketCap API key (optional, reserved for future CMC tools) |
MCP_HOST |
0.0.0.0 |
Server bind address |
MCP_PORT |
8000 |
Server port |
DATABASE_URL |
keys.db |
Path to the SQLite database |
CRYPTOSENSE_ENABLE_KEYGEN |
— | Set to true to expose the create_api_key admin tool |
Error Handling
All tools return a friendly {"error": "..."} dict on failure — no stack traces are ever returned to the client. Handled conditions:
- Invalid / missing API key → prompt to generate one
- Coin not found → suggests using the full CoinGecko ID
- Rate limit (429) → asks to wait and retry
- Network errors → descriptive message
- Invalid parameters → caught before the API call
License
MIT
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。