Overview
The VecTrade MCP server currently provides 42 financial and trading tools to any AI IDE that supports the Model Context Protocol. Supported clients: Claude Code, Claude Desktop, Cursor, VS Code Copilot, Windsurf, Cline, Continue.dev.Quick Setup
Option 1: Hosted (recommended — zero install)
Add this to your IDE’s MCP configuration:Option 2: Local (stdio via npx)
CLI Auto-Setup
If you have the VecTrade CLI:Claude Code: Plugin, Tools, MCP, and Skills
For Claude Code, treat the MCP server as your “plugin” layer:- Add VecTrade MCP server (hosted endpoint or local
npxserver). - Provide your user API key (
vq_...) for market-data tools. - Optionally add Bot key (
tvt_...) for trading tools. - Validate setup with
vectrade mcp doctor.
Recommended (hosted) for Claude Code
Optional Claude Code skill file
If your team uses Claude Code skill prompts, add a repo-level skill such as.claude/skills/vectrade/SKILL.md (template available in the VecTrade MCP repo) with tool-intent patterns:
Claude CLI Marketplace Install (one-time)
If you use Claude CLI and want a one-command plugin install path:vectrade-mcp-pluginskill namespace for consistent trading prompts- plugin-bundled
.mcp.jsonconnector (localnpx @vectrade/mcp-server)
Works Across Antigravity, VS Code, Cursor, and Claude Code
Yes. The same approach works for all MCP-compatible clients:- Use the same hosted MCP endpoint:
https://mcp.vectrade.io/mcp - Pass a user API key in header:
X-API-Key: vq_... - Keep a stable server name (recommended:
vectrade) - Restart the client so it reloads MCP config
- Verify by calling one tool (for example,
get_quote)
Client-specific config shapes
Different clients use slightly different JSON keys (mcpServers vs servers).
Config file locations by OS
Use this table to find where to place each config snippet:If Antigravity exposes both global and workspace MCP config, prefer workspace config for team consistency and easier onboarding.
url, headers, X-API-Key) and adapt only the wrapper key expected by that client version.
End-to-End User Guide
1) Get API key
- Go to vectrade.io/vtrade/developer
- Create or copy your key (
vq_...)
2) Add MCP server config
- Pick the client block above (Claude Code, Cursor, VS Code, Antigravity)
- Add
vectradeserver with endpointhttps://mcp.vectrade.io/mcp - Set
X-API-Key
3) Optional trading tools
- Add
VECTRADE_BOT_KEY=tvt_...only when you need trading execution tools - Keep market-data key (
vq_...) and bot key (tvt_...) separate
4) Restart client and verify
- Restart IDE/client
- Ask:
Use get_quote for AAPL - Confirm tool call returns data
5) Run doctor checks (recommended)
Troubleshooting
If issues persist, include
vectrade mcp doctor output when opening support requests.
Available Tools
Quotes & Market Data
Fundamentals
Technical Analysis
News & Sentiment
Analyst
Earnings & Insider
Options & ETF
Trading (Bot API)
Trading tools require a Bot API key (
tvt_...) set in VECTRADE_BOT_KEY. See the Bot Trading guide for setup.Trading Setup
To enable AI agent trading, add your Bot API key:Example Usage
Once configured, ask your AI assistant:- “What’s Apple’s current stock price?”
- “Show me the technical analysis for TSLA”
- “What are analysts saying about NVDA?”
- “Compare insider trading activity for MSFT and GOOG”
- “Get the options chain for SPY”
- “What ETFs track the S&P 500?”
- “Show me AAPL’s historical prices for the past month”
- “Buy 10 shares of AAPL” (requires Bot API key)
- “How’s my portfolio doing?” (requires Bot API key)
- “Set a stop-loss on TSLA at $250” (requires Bot API key)
Verification
Test the connection by asking your AI assistant to call any tool:“Use the get_quote tool to get the price of AAPL”If authentication fails, verify your API key is correct and active at vectrade.io/vtrade/developer.
Links
- GitHub Repository
- Examples
- Custom AI Agent Examples — Build your own trading agent (OpenAI, Claude, cron)
- Get API Key