near-liquid-staking
A Model Context Protocol (MCP) server for interacting with liquid staking protocols on NEAR blockchain. Supports Meta Pool (stNEAR) and Linear Protocol (LiNEAR).
README
MCP NEAR Liquid Staking Server
A Model Context Protocol (MCP) server for interacting with liquid staking protocols on NEAR blockchain. Supports Meta Pool (stNEAR) and Linear Protocol (LiNEAR).
Features
- stake_near - Stake NEAR tokens to receive liquid staking tokens (stNEAR or LiNEAR)
- unstake - Request unstaking of liquid tokens (starts the unbonding period)
- withdraw - Withdraw NEAR after the unstaking period completes
- get_staking_info - Query a user's staking position, balances, and pending withdrawals
- get_pool_stats - Get pool statistics including APY, TVL, exchange rate, and fees
- list_validators - List validators used by each staking pool
Supported Protocols
| Protocol | Contract | Token | Typical APY |
|---|---|---|---|
| Meta Pool | meta-pool.near |
stNEAR | ~9-10% |
| Linear | linear-protocol.near |
LiNEAR | ~10% |
Installation
npm install
npm run build
Usage
As MCP Server
Start the server:
npm start
# or
node dist/index.js
The server communicates over stdio using the MCP protocol.
MCP Configuration
Add to your MCP client configuration:
{
"mcpServers": {
"near-liquid-staking": {
"command": "node",
"args": ["/path/to/mcp-near-liquid-staking/dist/index.js"]
}
}
}
Tools
stake_near
Stake NEAR tokens on Meta Pool or Linear.
{
"protocol": "meta-pool",
"amount": "100",
"accountId": "your-account.near"
}
Returns transaction data ready for signing. Estimated liquid tokens to receive are calculated based on current exchange rate.
unstake
Request unstaking. Initiates the unbonding period (~48 hours / 4 epochs).
{
"protocol": "linear",
"amount": "50",
"accountId": "your-account.near"
}
withdraw
Withdraw NEAR after the unstaking period completes.
{
"protocol": "meta-pool",
"accountId": "your-account.near"
}
get_staking_info
Get user's complete staking position.
{
"protocol": "meta-pool",
"accountId": "your-account.near"
}
Response includes:
stakedBalance- Total NEAR equivalent stakedliquidTokenBalance- stNEAR/LiNEAR balancependingWithdrawal- Amount in unbondingwithdrawableBalance- Ready to withdrawepochsUntilWithdrawable- Epochs remaining until withdrawal
get_pool_stats
Get pool statistics and current rates.
{
"protocol": "linear"
}
Response includes:
totalStaked- Total NEAR staked in poolexchangeRate- Current NEAR per liquid token rateapy- Annual percentage yieldtvlNear- Total value lockedfee- Protocol fee percentagevalidatorCount- Number of validators
list_validators
List validators used by the staking pool.
{
"protocol": "meta-pool"
}
Development
# Install dependencies
npm install
# Run in development mode
npm run dev
# Run tests
npm test
# Build
npm run build
Testing
npm test
Tests cover:
- Pool stats retrieval for both protocols
- Staking info queries
- Stake/unstake/withdraw transaction preparation
- Validator listing
- Edge cases (small/large amounts)
Architecture
src/
├── index.ts # MCP server entry point with tool handlers
└── near-client.ts # NEAR RPC client for staking operations
The server uses:
@modelcontextprotocol/sdk- MCP server implementationnear-api-js- NEAR blockchain interactionzod- Input validation
Notes
- Transaction Signing: This server prepares transaction data but does not sign transactions. The consuming application must handle wallet integration and signing.
- RPC Endpoints: Uses public NEAR RPC endpoints. For production, consider using a dedicated RPC provider.
- Exchange Rates: Liquid token amounts are calculated based on current exchange rates which may vary slightly at execution time.
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。
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。
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。
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。