BYTE Protocol
Per-byte data marketplace for AI agents on Arbitrum. Discover publishers, evaluate on-chain Proof-of-Quality Score (PQS), subscribe + pay per request in USDC via the x402 gateway. 13 tools, no API keys, live testnet.
README
Byte Protocol MCP Server
An MCP (Model Context Protocol) server that gives AI agents direct access to the Byte Protocol data marketplace on Arbitrum Sepolia.
What is MCP?
The Model Context Protocol is an open standard that lets AI assistants (like Claude) use external tools. This server exposes Byte Protocol's on-chain operations as MCP tools, so an AI agent can discover data publishers, check reputation scores, subscribe to feeds, and publish data -- all through natural language.
Tools
This server exposes 10 tools across read and write operations:
Read-only (no wallet required)
| Tool | Description |
|---|---|
byte_search_publishers |
Search publishers by topic, minimum PQS score, and sort order |
byte_get_publisher |
Get detailed on-chain info for a publisher (status, tier, stake, PQS breakdown) |
byte_get_network_stats |
Get network-wide stats (total publishers, messages, fees, revenue) |
byte_check_subscription |
Check if an address is subscribed to a specific publisher |
byte_get_token_balances |
Get PPB, USDC, and ETH balances for any address |
byte_list_feeds |
List all active data feeds with pricing and quality scores |
Write (require PRIVATE_KEY)
| Tool | Description |
|---|---|
byte_drip_faucet |
Request 500 testnet PPB tokens (24h cooldown, 1000 PPB lifetime cap) |
byte_subscribe |
Subscribe to a publisher's data feed |
byte_register_publisher |
Register as a data publisher (schema + stake + on-chain registration) |
byte_publish_data |
Publish data to a subscriber via the DataStream contract |
Installation
git clone https://github.com/byte-protocol/mcp-server.git
cd mcp-server
npm install
npm run build
Configuration
Claude Desktop
Add this to your Claude Desktop config (~/.config/claude/claude_desktop_config.json on Linux, ~/Library/Application Support/Claude/claude_desktop_config.json on macOS):
{
"mcpServers": {
"byte-protocol": {
"command": "node",
"args": ["/path/to/mcp-server/dist/index.js"],
"env": {
"RPC_URL": "https://sepolia-rollup.arbitrum.io/rpc",
"INDEXER_URL": "http://localhost:8080"
}
}
}
}
To enable write operations (subscribe, publish, faucet), add your private key:
{
"mcpServers": {
"byte-protocol": {
"command": "node",
"args": ["/path/to/mcp-server/dist/index.js"],
"env": {
"PRIVATE_KEY": "0x...",
"RPC_URL": "https://sepolia-rollup.arbitrum.io/rpc",
"INDEXER_URL": "http://localhost:8080"
}
}
}
}
Claude Code
claude mcp add byte-protocol node /path/to/mcp-server/dist/index.js
Environment Variables
| Variable | Required | Default | Description |
|---|---|---|---|
PRIVATE_KEY |
No | -- | Wallet private key for write operations. Without it, only read tools work. |
RPC_URL |
No | https://sepolia-rollup.arbitrum.io/rpc |
Arbitrum Sepolia RPC endpoint |
INDEXER_URL |
No | http://localhost:8080 |
Byte Protocol indexer API URL |
Usage
Once connected, your AI agent can do things like:
"Search for weather data publishers with a PQS score above 5000"
"What are the network stats for Byte Protocol?"
"Check the balance of 0xABC...123 on the testnet"
"Subscribe me to publisher 0xDEF...456"
"Register as a publisher for the topic eth-price with a 50 PPB stake"
"Get me some testnet tokens from the faucet"
Development
npm run dev # Watch mode -- recompiles on changes
npm run build # One-time build
npm start # Run the server
Network
This server connects to Arbitrum Sepolia (chain ID 421614) by default. All contract addresses are for the testnet deployment.
License
MIT -- see LICENSE.
Links
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。