DROYD MCP Server
# DROYD MCP Server
A Model Context Protocol (MCP) server for the [DROYD](https://droyd.ai) AI trading agent API. This server enables Claude Desktop, Claude Code, and other MCP clients to interact with DROYD's crypto research, project discovery, and trading capabilities.
## Features
- **Agent Chat** - Multi-turn conversations with DROYD AI agents (research, trading, data)
- **Content Search** - Search DROYD's knowledge base for crypto news, posts, tweets, and more
- **Project Discovery** - Find projects by name, symbol, address, or semantic concepts
- **Market Screening** - Filter projects using market criteria or natural language
- **Watchlists** - Access personal and swarm watchlists with agent evaluations
- **Trading** - Open positions with flexible leg configurations (stops, takes, etc.)
- **Position Management** - Manage trades and view P&L
## Installation
### Option 1: npm (Recommended)
```bash
npx @droyd/mcp-server
```
### Option 2: Clone Repository
```bash
git clone https://github.com/droyd/mcp-server.git
cd mcp-server
npm install
npm run build
```
## Configuration
### Claude Desktop
Add to your `claude_desktop_config.json`:
```json
{
"mcpServers": {
"droyd": {
"command": "npx",
"args": ["-y", "@droyd/mcp-server"],
"env": {
"DROYD_API_KEY": "your-api-key"
}
}
}
}
```
### Claude Code
Add to your MCP settings:
```json
{
"mcpServers": {
"droyd": {
"command": "npx",
"args": ["-y", "@droyd/mcp-server"],
"env": {
"DROYD_API_KEY": "your-api-key"
}
}
}
}
```
### Local Development
```json
{
"mcpServers": {
"droyd": {
"command": "node",
"args": ["/path/to/droyd-mcp-server/dist/index.js"],
"env": {
"DROYD_API_KEY": "your-api-key"
}
}
}
}
```
## Environment Variables
| Variable | Required | Description |
|----------|----------|-------------|
| `DROYD_API_KEY` | Yes | Your DROYD API key |
| `DROYD_BASE_URL` | No | Override API base URL (default: https://api.droyd.ai) |
## Available Tools
### Agent
| Tool | Description |
|------|-------------|
| `droyd_chat` | Chat with a DROYD AI agent (research, trading, data). Supports multi-turn conversations. |
### Search
| Tool | Description |
|------|-------------|
| `droyd_search_content` | Search the knowledge base for posts, news, tweets, YouTube, developments |
### Projects
| Tool | Description |
|------|-------------|
| `droyd_search_projects` | Find projects by name, symbol, address, or semantic query |
| `droyd_filter_projects` | Screen projects using market filters or natural language |
| `droyd_get_watchlist` | Get personal/swarm watchlists with agent evaluations |
### Trading
| Tool | Description |
|------|-------------|
| `droyd_open_trade` | Open positions with flexible leg configurations |
| `droyd_manage_trade` | Close, buy, sell, or update existing positions |
| `droyd_get_positions` | View active positions, wallet holdings, and P&L |
## Development
```bash
# Install dependencies
npm install
# Build
npm run build
# Run in development mode
npm run dev
# Type check
npm run typecheck
# Lint
npm run lint
```
## API Reference
For detailed API documentation, see [DROYD API Docs](https://docs.droyd.ai).
### Rate Limits
- 10 requests per 15 minutes per endpoint
- HTTP 429 returned when limit exceeded
## License
MIT
TDQS
Scored across 8 tools
Most tools have distinct purposes: chat, filter projects, get positions, manage trades, open trades, search content, and search projects are clearly separated. However, droyd_get_positions and droyd_get_watchlist could be slightly overlapping in providing portfolio-related data, but their descriptions clarify that positions are for active trades while watchlists are for tracked projects.
All tools follow a consistent 'droyd_verb_noun' pattern with snake_case throughout: droyd_chat, droyd_filter_projects, droyd_get_positions, droyd_get_watchlist, droyd_manage_trade, droyd_open_trade, droyd_search_content, droyd_search_projects. This makes them predictable and easy to identify.
With 8 tools, this is well-scoped for a crypto trading and research server. Each tool serves a specific function in the domain, from chatting with AI agents to managing trades and searching for projects or content, without being overwhelming or too sparse.
The toolset covers core crypto workflows: research (chat, search), project discovery (filter, search), portfolio management (get positions, watchlist), and trading (open, manage). A minor gap is the lack of a tool for direct market data queries or analytics beyond chat, but the chat tool with agent types like 'data' and 'trading' can partially fill this role.