stock-market-analysis

This project is a Spring Boot WebFlux application that uses Spring AI with the Google Gemini Developer API.

It includes two simple agents:

• Research Agent: gathers notes for the user request and can use Yahoo Finance MCP tools when enabled.
• Signal Agent: turns those notes into a final BUY/SELL-style signal response.

• Java 21+
• Gradle 8+
• A Gemini API key from Google AI Studio
• uv / uvx if you want to run the Yahoo Finance MCP server locally

Set these environment variables before running:

export GOOGLE_API_KEY=your-google-api-key
export GEMINI_MODEL=gemini-2.0-flash

Optional Yahoo Finance MCP setup:

export MCP_YAHOO_ENABLED=true
export MCP_YAHOO_COMMAND=uvx
export MCP_YAHOO_PACKAGE=mcp-yahoo-finance

When enabled, Spring AI starts the mcp-yahoo-finance MCP server over STDIO and exposes its tools to the research agent.

gradle bootRun

GET request:

curl "http://localhost:8080/api/agents/run?task=HAL india"

POST request:

curl -X POST "http://localhost:8080/api/agents/run" \
  -H "Content-Type: application/json" \
  -d '{"task":"Summarize Tesla stock context and turn it into an investor-friendly brief"}'

The app also includes a browser UI at:

http://localhost:8080

• stock to analyze: input box for the stock prompt/task.
• Run with GET: runs workflow via GET /api/agents/run?task=....
• Run with POST: runs workflow via POST /api/agents/run.
• Clear Output: clears current research/signal output areas.
• Research Agent Output: raw/structured research collected by the research agent.
• Signal Agent Output: final recommendation text from the signal agent.
• MCP Tools In Use: loaded MCP tools from GET /api/agents/mcp/tools.

• Analyze Indian Market: runs `GET /api/agents/india/