stock-intelix
# Stock Intelix — Model Context Protocol (MCP) Server
[](https://modelcontextprotocol.io)
[](https://opensource.org/licenses/MIT)
[](https://nodejs.org/)
Official Model Context Protocol (MCP) server for [Stock Intelix](https://stockintelix.com) — quantitative equity intelligence and probabilistic price range forecasting across 34 global exchanges.
Designed for seamless integration with **Claude Desktop**, **Cursor**, **Glama**, and any AI agent supporting the [Model Context Protocol](https://modelcontextprotocol.io).
---
## Capabilities & Tools
This MCP server equips AI assistants with two specialized financial tools:
### 1. `get_forecast`
Retrieve quantitative probabilistic price forecasts and risk horizons for any covered equity ticker.
- **Parameters:**
- `symbol` *(string, required)*: Equity ticker (e.g., `AAPL`, `NVDA`, `RELIANCE.BSE`, `SAP.XETRA`).
- `horizon` *(integer, optional)*: Forecast horizon in trading days from 5 to 120 (default: 20).
- **Output:** Current price, directional probability (bullish/bearish odds), probabilistic percentiles (p10, p25, median p50, p75, p90), calibration metrics, and catalyst risk score.
### 2. `search_securities`
Search the Stock Intelix global database across 34 exchanges to find the exact symbol for any company.
- **Parameters:**
- `query` *(string, required)*: Company name or ticker fragment (min 2 characters, e.g. `Microsoft` or `MSFT`).
- `exchange` *(string, optional)*: Filter by market code (e.g. `NASDAQ`, `NYSE`, `BSE`, `LSE`, `DFM`).
- **Output:** List of matching equities, exchange codes, and operational status.
---
## Authentication & API Keys
All calls to the Stock Intelix API require an API key:
1. Sign up or log in at [stockintelix.com](https://stockintelix.com).
2. Navigate to **Console / Developers** (`https://stockintelix.com/app/developers`).
3. Generate an API Key (starts with `sik_...`).
Set the API key as an environment variable:
```bash
export STOCK_INTELIX_API_KEY="sik_your_api_key_here"
```
---
## Quickstart
### Option A: Using with Claude Desktop
Add this configuration to your `claude_desktop_config.json`:
```json
{
"mcpServers": {
"stock-intelix": {
"command": "npx",
"args": ["-y", "stock-intelix-mcp"],
"env": {
"STOCK_INTELIX_API_KEY": "sik_your_api_key_here"
}
}
}
}
```
### Option B: Using with Cursor
1. Open Cursor **Settings** → **Features** → **MCP**.
2. Click **+ Add New MCP Server**.
3. Select **Command / Stdio**:
- **Name:** `stock-intelix`
- **Command:** `npx -y stock-intelix-mcp`
- **Environment Variable:** `STOCK_INTELIX_API_KEY=sik_your_api_key_here`
### Option C: Using Docker / Glama
Build and run using the included Dockerfile:
```bash
# Build container
docker build -t stock-intelix-mcp .
# Run with your API key
docker run -i --rm -e STOCK_INTELIX_API_KEY="sik_your_api_key_here" stock-intelix-mcp
```
---
## Development & Local Testing
```bash
# Clone the repository
git clone https://github.com/stockintelix/stock-intelix-mcp.git
cd stock-intelix-mcp
# Install dependencies
npm install
# Test run locally
STOCK_INTELIX_API_KEY="sik_test" node index.js
```
---
## Disclaimer
All forecasts and quantitative metrics provided by Stock Intelix are informational models and data analytics only. They do not constitute financial advice, buy/sell recommendations, or endorsements.
---
## License
[MIT](LICENSE) © 2026 Stock Intelix
TDQS
Scored across 2 tools
The two tools have clearly distinct purposes: search_securities finds securities by name or ticker, while get_forecast retrieves forecasts for a known ticker. There is no meaningful overlap, so an agent should not confuse them.
Both names use the same snake_case verb_noun convention: get_forecast and search_securities. There is no mixing of styles or abbreviations.
Only two tools are exposed, which is thin even for a focused forecast/search server. They are well chosen, but the surface feels minimal rather than fully scoped.
The core read-only workflow is covered: find a security by name or ticker, then retrieve a probabilistic forecast. Minor gaps remain around batch comparisons or exchange metadata, but there is no obvious dead end for the stated purpose.