
Seitrace Insights MCP Server
Enables natural language access to blockchain data and insights through the Seitrace API. Supports querying address details, token information (ERC-20, ERC-721, ERC-1155, CW20, CW721), transaction data, and smart contract details across multiple blockchain networks.
README
<img src="https://assets.seitrace.com/seitrace-mcp.gif?v=10" width="1280"></img>
Seitrace Insights MCP Server 🚀
A Model Context Protocol (MCP) server that exposes the Seitrace Insights API as model-friendly tools. It now advertises five focused tools that implement a resource-based interface so LLMs can discover resources, list actions, fetch schemas, invoke them, and generate code snippets.
Highlights ✨
What MCP provides to end users and assistants:
- Natural‑language access to Seitrace insights. The assistant performs API calls on your behalf.
- Self‑describing tool flow: enumerate actions, retrieve the input schema, then invoke.
- Input validation and clear error messages using per‑action JSON Schemas.
- Concise discovery: minimal list output; detailed payloads only when invoking actions.
- Integration with MCP‑enabled VS Code extensions (e.g., Continue, Cline).
- Simple, secure API key handling via environment variables (sent as
x-api-key
). - Quick start via npx:
npx -y @seitrace/mcp
.
Getting started
Make sure you obtain an API Key for free here
Use with VSCode variants, Claude Desktop / Cursor 💻
See mcp
Using with an MCP Client 🤝
Configure your MCP client to launch the compiled server binary:
- Command:
npx
- Args:
["-y", "@seitrace/mcp"]
- Env:
SECRET_APIKEY
,API_BASE_URL
(optional)
Once connected, the client will call tools/list
, which returns exactly five tools representing the resource interface.
Available tools 🧰
Primary entrypoint: five tools that form the resource-based interface:
list_resources
— list available resourceslist_resource_actions
— list actions for a resourceget_resource_action_schema
— get the JSON Schema for an actioninvoke_resource_action
— invoke an action with payloadget_resource_action_snippet
— generate a code snippet to perform a resource action in the specified language, for example, a javascript snippet to call the action with the required parameters
Flow: list resources -> list actions for a resource -> get action schema -> invoke action or generate a code snippet.
Common resources include:
Insights
insights_address
— address detail, transactions, and token transfers.insights_erc20
— ERC‑20 token information, balances, transfers, and holders.insights_erc721
— ERC‑721 token information, transfers, and holders.insights_erc1155
— ERC‑1155 token information, instances, and holders.insights_cw20
— CW20 token information, balances, transfers, and holders.insights_cw721
— CW721 token information, transfers, and holders.insights_ics20
— ICS‑20 (IBC fungible) transfer information.insights_native
— native token information and statistics.insights_smart_contract
— smart contract detail.
General
general_faucet
- enable requesting faucet for developers
Typical Flow 🔁
Using the MCP SDK, drive the resource-based flow via the five tools:
// 1) Discover available resources
const resources = await client.callTool({ name: 'list_resouces', arguments: {} });
// -> { resources: ['erc20', 'erc721', 'native', ...] }
// 2) List actions for a resource
const actions = await client.callTool({ name: 'list_resouce_actions', arguments: { resource: 'insights_erc20' } });
// -> { resource: 'erc20', actions: [{ name, description }, ...] }
// 3) Get the JSON Schema for a specific action
const schema = await client.callTool({ name: 'get_resource_action_schema', arguments: { resource: 'insights_erc20', action: 'get_erc20_token_info' } });
// -> { resource: 'erc20', action: 'get_erc20_token_info', schema }
// 4) Invoke the action with its payload
const res = await client.callTool({ name: 'invoke_resource_action', arguments: { resource: 'insights_erc20', action: 'get_erc20_token_info', payload: { chain_id: 'pacific-1', contract_address: '0x...' } } });
// res.content[0].text -> "API Response (Status: 200):\n{ ... }"
// 5) Optionally, generate a code snippet for an action
const snippet = await client.callTool({ name: 'get_resource_action_snippet', arguments: { resource: 'insights_erc20', action: 'get_erc20_token_info', language: 'node' } });
// -> { resource, action, language, snippet }
The server validates payload
against the action’s schema and returns a pretty-printed JSON body when applicable.
Requirements 🔧
- Node.js 20+
- A Seitrace Insights API key (optional for discovery, required for most live calls), obtain it here
Install 📦
npm install
Configure 🔐
Copy .env.example
to .env
and set your values as needed.
Environment variables:
API_BASE_URL
(optional) — defaults tohttps://seitrace.com/insights
SECRET_APIKEY
— Seitrace API key; used to set headerx-api-key
Build and Run 🏃
# Type-check and compile to build/
npm run build
# Run the MCP server over stdio (used by MCP clients)
npm start
This server is designed to be launched by an MCP-compatible client (e.g., via a command/args configuration). It communicates over stdio.
End-to-End Test ✅
Run the E2E to verify the root resource flow and (optionally) a live positive-call:
# Optionally provide your API key so the positive path runs
SEITRACE_API_KEY=your_key_here npm run test:e2e
Troubleshooting 🛠️
- Validation errors: If
invokeResourceAction
returns “Invalid arguments…”, callgetResourceActionSchema
and ensure yourpayload
follows the schema. - Unknown action: You’ll get an error that includes the available actions. Use
listResourceActions
to discover the right name. - 401/403 responses: Set
SECRET_APIKEY
with a valid Seitrace key. - Network issues: Ensure
API_BASE_URL
is reachable from your environment. - Node version: Use Node 20+ as required in
package.json
.
Contributing 🤝
- Keep
tools/list
output compact. Do not embed per-action details there—fetch them viagetResourceActionSchema
. - New endpoints should appear under the correct resource; root tool methods should provide discovery and invocation consistently.
- Prefer small, focused modules in
src/lib/
for shared logic.
License 📄
See LICENSE
Support 📨
Please shoot emails to dev@cavies.xyz
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。