“Find alpha-positive small-caps with insider buying in the last 30 days.”
Screens the universe, cross-checks insider filings, ranks by edge, returns the names with their backtests.
TraderGenie runs a hosted MCP server. Connect Claude Code, Claude Desktop, or Cursor with a URL and a key — no clone, no SDK, no local process. 100+ tools across 13 capability groups, all driven from plain language.
“Find alpha-positive small-caps with insider buying in the last 30 days.”
Screens the universe, cross-checks insider filings, ranks by edge, returns the names with their backtests.
“Build a mean-reversion strategy from this idea, walk-forward validate it, deploy if it survives.”
Generates the Python, runs the train/test split, checks the overfit ratio, and deploys only if it passes.
“How is AAPL set up right now?”
Pulls price, fundamentals, signal convergence, news and social sentiment into one read.
Generate a key at Settings → MCP Server Access on app.tradergenie.ai.
Claude Code — one command:
claude mcp add --transport http tradergenie \
https://app.tradergenie.ai/mcp \
--header "X-Api-Key: tg_your_key_here" or drop this into .mcp.json (Claude Desktop / Cursor / any client):
{
"mcpServers": {
"tradergenie": {
"type": "http",
"url": "https://app.tradergenie.ai/mcp",
"headers": { "X-Api-Key": "tg_your_key_here" }
}
}
} Your key is your identity — every tool call runs as your TraderGenie user. Revoke it anytime.
The same engine the dashboards use — exposed as tools you can call directly.
Screen the whole universe by signal, then stack conditions until only the names that matter survive.
screen_signals Universe screen with 15+ gates — GARP, market-cap bounds, valuation, growth, analyst rating, anti-chase thresholds — in one call. screen_convergence Multi-signal AND logic (up to 12 conditions) with cohort routing and earnings blackout — only symbols where the strategies agree. get_signal_methodology How any signal works, plus its backtest and training metrics. Every strategy carries its trust label — and you can recompute it yourself, the way an analyst would.
research_strategy One-call verdict: cross-sectional validation + FDR correction + correlations + rolling walk-forward. get_walk_forward_results Per-symbol overfit detection on a train/test split — validated, caution, likely-overfit, insufficient data. compute_fdr_correction False-discovery-rate gate so multiple testing doesn't hand you a fluke. Describe a strategy in plain English, get Python, backtest it, deploy it across the universe.
generate_script Natural-language → trading-signal code (score, strategy, indicator, or metric). deploy_script Turn a script into a recurring subscription — scheduled, realtime, or manual. run_batch Execute across a watchlist or the full universe. Define, train, and deploy your own predictive models — their output becomes signals like any other.
create_ai_model Define a model: type, prediction target, symbol scope, features. deploy_ai_model Publish a trained model's predictions as live signals. get_inference_results Latest predictions per symbol. The full pick pipeline and risk-budgeted portfolio build — with an auditable reason for every dismissal.
run_strategy_funnel Multi-stage pick funnel — pre-flight → sector rotation → convergence → edge → AI ensemble → validation → tier — with an audit table and dismissal log. build_portfolio Convergence + risk budget + next steps per candidate. get_monitor_snapshot Weekly read: breadth, sector rotation, Fama-French, VIX regime, data gaps. Prices, factors, breadth, and sector rotation across the whole universe.
get_price_bars Daily OHLCV for trend analysis and charting. get_sector_rotation Leading and lagging sectors by return. get_fama_french_factors Factor exposure — value, size, profitability, momentum. Deep fundamentals and a multi-field screener, batched across hundreds of symbols.
screen_fundamentals Screen on P/E, margins, growth, debt, beta, analyst rating, and more. compare_symbols Side-by-side fundamentals + signals across 2–10 stocks. Cluster-buy insider signals, activist filings, and outlier-compounder quality scoring.
get_insider_candidates Rank by insider metrics — cluster buys, net dollar buying, activist 13D filings. get_compounder_candidates Rank by outlier compounder score — growth, profitability, FCF yield, ROE, low debt. Decompose returns, size positions with Kelly, and check portfolio concentration.
analyze_portfolio_risk VaR / CVaR, correlations, and sector concentration. get_factor_decomposition Split a stock's returns into market beta, sector beta, and alpha. get_hit_rate BUY-signal outcome distribution and a Kelly fraction for sizing. Headlines and aggregated retail sentiment, scored and filtered to your tickers.
get_news_sentiment Headlines with a sentiment score and a running average. get_social_sentiment StockTwits / Reddit / ApeWisdom with FinBERT scores. get_trending_tickers Social-mention leaderboard. Create and read watchlists and holdings — the context everything else binds to.
create_watchlist New watchlist, optionally seeded with symbols. get_holdings Portfolio holdings with position data. Compose dashboards from ready-made templates and 25+ widget types — entirely from the conversation.
create_dashboard Build a dashboard from a ready-made intent template. add_widget Place any of 25+ widget types — KPI, signal chart, screener, news, heatmap, backtester, and more. Wire alerts on any signal and manage the scheduled jobs that compute them.
create_signal_change_alert Fire on BUY / SELL / NEUTRAL transitions on any deployed signal. create_alert Threshold alert with notification routing. list_subscriptions See and manage deployed subscriptions and their schedules. MCP access ships with your invite. Be among the first to drive the platform from your editor.
Join the waitlist →