vfat Metrics MCP
Enables querying public vfat Metrics position data by EVM address, providing structured access to current positions, NFT lifecycle actions, historical value/performance, CSV exports, and portfolio summaries through MCP tools.
README
vfat Metrics MCP
A read-only Model Context Protocol server for querying public vfat Metrics position data by EVM address. It gives MCP clients structured access to current vfat/Sickle positions, NFT lifecycle actions, historical value and performance series, CSV exports, and a compact portfolio summary.
This is an independent community project. It is not affiliated with or endorsed by vfat, Sickle, or their contributors.
Tools
| Tool | Purpose |
|---|---|
get_vfat_positions |
Return every current position and the provider's position fields |
get_vfat_position |
Find one position by NFT/token ID |
get_vfat_portfolio_summary |
Count positions by chain/protocol and aggregate selected USD metrics |
get_vfat_position_actions |
Group deposits, rebalances and other events by transaction |
get_vfat_position_value_history |
Return historical value, token balances and range state |
get_vfat_position_performance |
Return APR, ROI, PnL and balance time series |
export_vfat_position_csv |
Export actions, value history or performance as CSV text |
Every tool is declared read-only and idempotent. The server cannot sign messages, submit transactions, rebalance liquidity, or move funds.
Requirements
- Node.js 20 or newer
- An MCP client such as Codex, Claude Desktop, or another MCP-compatible agent
- Internet access to
https://info-api.vf.at
No API key is required.
Install and run
From a checkout:
npm install
npm run build
node dist/index.js
For development:
npm run dev
The default transport is stdio. The process writes protocol messages to stdout and diagnostics to stderr.
Codex configuration
Build the project, then add it to ~/.codex/config.toml:
[mcp_servers.vfat_metrics]
command = "node"
args = ["/absolute/path/to/vfat-metrics-mcp/dist/index.js"]
Restart Codex after changing MCP configuration. You can then ask, for example:
Show and summarize the current vfat positions for
<EVM_ADMIN_ADDRESS>.
Analyze rebalance costs and cashflows for the position identified by
<CHAIN_ID>,<SICKLE_ADDRESS>,<NFT_ID>, and<NFT_MANAGER_ADDRESS>.
Claude Desktop configuration
{
"mcpServers": {
"vfat-metrics": {
"command": "node",
"args": ["/absolute/path/to/vfat-metrics-mcp/dist/index.js"]
}
}
}
Streamable HTTP
For local development or a trusted private network:
npm run build
node dist/index.js --http
- MCP endpoint:
http://127.0.0.1:3000/mcp - Health endpoint:
http://127.0.0.1:3000/health
HTTP mode is stateless. Do not expose it publicly without authentication, TLS, request limits, and
a suitable reverse proxy. Configure it with MCP_HTTP_HOST, MCP_HTTP_PORT, or
MCP_TRANSPORT=http.
Data and precision
The server reads public endpoints used by the vfat Metrics website:
GET https://info-api.vf.at/open-positions-v2?admin_address=<address>
GET https://info-api.vf.at/sickle-nft-actions?...position identity...
POST https://info-api.vf.at/historical-underlying-assets
GET https://info-api.vf.at/position-performance-history?...position identity...
The upstream endpoint is not documented as a stable public API and may change without notice. Raw
position fields are passed through so new fields remain available. Portfolio totals use
decimal.js; provider-supplied decimal strings are therefore summed without JavaScript binary
floating-point addition.
USD values, PnL, ROI, APR, token prices, timestamps, range state, and rewards are provider-derived estimates. They may be delayed, incomplete, or calculated differently from on-chain state. Verify material decisions independently with on-chain reads and protocol contracts.
Position history
vfat can emit several raw rows for one transaction. get_vfat_position_actions groups them by
transaction hash, selects the lifecycle action, sums provider cashflow fields with decimal
arithmetic, and retains rewards and swaps. Set include_raw_events=true when an audit needs every
provider row.
Value history follows NFT migrations and rebalances by discovering every token ID in the action
chain before requesting historical underlying assets. Points at the same timestamp are aggregated
into normalized token balances, USD values, and one in_range state.
All history tools accept optional inclusive from and to ISO 8601 timestamps. Position identity
fields are supplied at runtime and are never stored by the server.
CSV export
export_vfat_position_csv accepts dataset=actions, value_history, or performance. It returns
UTF-8 CSV text plus structured metadata (filename, mime_type, and row_count). The CSV is built
from the same normalized JSON returned by the history tools, making calculations reproducible
without browser automation.
Configuration
Copy .env.example values into your process environment if you need overrides:
| Variable | Default | Meaning |
|---|---|---|
VFAT_METRICS_API_URL |
https://info-api.vf.at |
Upstream base URL |
VFAT_METRICS_TIMEOUT_MS |
15000 |
Request timeout in milliseconds |
MCP_TRANSPORT |
stdio |
stdio or http |
MCP_HTTP_HOST |
127.0.0.1 |
HTTP bind address |
MCP_HTTP_PORT |
3000 |
HTTP port |
Development
npm run check
This runs formatting checks, ESLint, TypeScript typechecking, unit tests, and a production build.
License
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。