Pump.fun MCP Server
A Cloudflare Worker that implements the MCP Streamable HTTP transport for AI agents to interact with pump.fun on Solana.
README
Pump.fun MCP Server
A Cloudflare Worker that implements the MCP Streamable HTTP transport for AI agents to interact with pump.fun on Solana.
Tools (10 read-only tools)
| Tool | Description |
|---|---|
searchTokens |
Search tokens by name, symbol, or mint address |
getTokenDetails |
Full details for a specific token |
getBondingCurve |
Bonding curve analysis — reserves, price, graduation % |
getTokenTrades |
Recent trade history (buys & sells) |
getTrendingTokens |
Top tokens by market cap |
getNewTokens |
Most recently launched tokens |
getGraduatedTokens |
Tokens that graduated to Raydium AMM |
getKingOfTheHill |
Highest market cap token still on bonding curve |
getCreatorProfile |
All tokens by a creator + rug-pull risk flags |
getTokenHolders |
Top holders with concentration analysis |
All tools are read-only — no wallet keys needed. Write operations (buy/sell/create) require client-side wallet signing and are intentionally excluded for safety.
Deploy
cd workers/pump-fun-mcp
npm install
npx wrangler deploy
Custom domain
To serve at pump-fun-sdk.modelcontextprotocol.name (or any subdomain):
- Add a CNAME record in Cloudflare DNS pointing to your Worker
- Add a Custom Domain in the Cloudflare dashboard under Workers → pump-fun-mcp → Settings → Domains & Routes
Or for the path-based gateway pattern (modelcontextprotocol.name/mcp/pump-fun-sdk):
- Uncomment the
[route]section inwrangler.toml - Redeploy
Connect from SperaxOS
The server is registered in cryptoMcpData.ts and appears in the MCP marketplace. Users can also add it manually:
https://pump-fun-sdk.modelcontextprotocol.name/mcp
Development
npx wrangler dev # local dev server on :8787
Test with curl:
# Discovery manifest
curl http://localhost:8787/
# Initialize
curl -X POST http://localhost:8787/mcp \
-H 'Content-Type: application/json' \
-d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2024-11-05","capabilities":{},"clientInfo":{"name":"test","version":"1.0"}}}'
# List tools
curl -X POST http://localhost:8787/mcp \
-H 'Content-Type: application/json' \
-d '{"jsonrpc":"2.0","id":2,"method":"tools/list","params":{}}'
# Search tokens
curl -X POST http://localhost:8787/mcp \
-H 'Content-Type: application/json' \
-d '{"jsonrpc":"2.0","id":3,"method":"tools/call","params":{"name":"searchTokens","arguments":{"query":"pepe","limit":5}}}'
Architecture
AI Agent (SperaxOS)
│
▼
Cloudflare Worker (MCP Streamable HTTP)
│
├── pump.fun API (token data, search, trades)
│ └── frontend-api-v3.pump.fun
│
└── Solana RPC (holder data, on-chain reads)
└── api.mainnet-beta.solana.com
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。