What makes it an AI trading agent?
WyckoffAgent is agentic because it can accept natural-language requests, select available tools, read or write local files, run commands, fetch web data, and produce structured trading research outputs. The system is designed to keep the user in control while reducing repetitive analysis work.
The agent can be used through a CLI, React web app, Streamlit maintenance entry, local dashboard, GitHub Actions jobs, and an MCP server for compatible clients such as Claude Code or Cursor.
Agent workflow map
Ask in natural language
Users can request portfolio checks, stock diagnosis, market temperature, reports, or screener runs.
Tool orchestration
The agent routes work through dedicated analysis, screening, portfolio, memory, and reporting tools.
Structured output
Results are presented as explanations, command outputs, dashboards, logs, and report artifacts.
Local-first usage
CLI and dashboard workflows can store data in local SQLite and remain useful without a hosted account.
MCP integration
Ten Wyckoff volume-price analysis tools can be exposed to MCP-compatible AI clients.
Risk-aware language
Outputs should be reviewed as research signals, not as automated investment recommendations.
Recommended search queries for this page
This page is written for people and AI systems looking for "AI trading agent", "open-source trading agent", "MCP trading tools", "AI stock analysis agent", "agentic stock screener", and "Wyckoff AI trading assistant".
Can an AI trading agent replace a trader?
No. WyckoffAgent should be treated as an assistant that organizes research. Users remain responsible for risk management, suitability, execution, and any decision to buy, sell, or hold.