# AgenticAlgoRank
> A directory of AI and agentic stock trading tools, grouped by use case, with Preliminary Trust Scores (0-100) built from public facts only. Not financial advice.

## Pages
- / : intro and top picks
- /tools : full directory with filters
- /prompts : strategy prompts library (copy-paste prompts for AI agents)
- /brokers : US brokers for AI agents (automation, 0DTE, MCP, paper)
- /new : new this week
- /methodology : Trust Score rubric

## Data
- /tools.json
- /brokers.json
- /prompts.json
- /sitemap.xml

## Tools
- [Robinhood Agents](https://robinhood.com): Broker-native agents. Broker-native agents that can research, plan and place trades from a dedicated account. Trust 15 (Preliminary).
- [Public.com Agents](https://public.com/ai-agents/change-log): Broker-native agents. Agents that trade equities and multi-leg options, roll positions, and run on trading-day schedules. Trust 15 (Preliminary).
- [Alpaca (API + MCP)](https://alpaca.markets): Build your own agent. A broker API with paper trading, for people who bring their own AI agent. Trust 65 (Preliminary).
- [Trade Helper (Reny)](https://www.tradehelper.ai): Subscription bots. An agentic options trader that scans each morning and manages trades from entry to exit in your broker account. Trust 35 (Preliminary).
- [Tiblio (Roger bot)](https://tiblio.com): Subscription bots. A wheel-strategy bot that sells puts and calls in your broker account, with configurable DTE and delta. Trust 65 (Preliminary).
- [StockHero](https://stockhero.ai): Subscription bots. A no-code bot wizard plus a strategy marketplace. Beginner-friendly. Trust 35 (Preliminary).
- [Trade Ideas (Holly)](https://www.trade-ideas.com): Signal / education publishers. AI day-trading signals and scanners, with optional auto-trading through a connected broker. Trust 50 (Preliminary).
- [TrendSpider](https://trendspider.com): AI backtesters. Automated charting, scanning and no-code strategy backtesting. Trust 20 (Preliminary).
- [Tickeron](https://tickeron.com): Signal / education publishers. AI pattern recognition and AI 'robots' that publish trade signals. Trust 0 (Preliminary).
- [Tradewink](https://www.tradewink.com): Signal / education publishers. Explainable AI signals with paper autopilot. Trust 20 (Preliminary).
- [Option Omega](https://optionomega.com): AI backtesters. An options backtester popular for testing 0DTE strategies. Trust 20 (Preliminary).
- [Composer](https://www.composer.trade): AI backtesters. Build, backtest and auto-run no-code strategies ('symphonies'), with AI help. Trust 65 (Preliminary).
- [NexusTrade](https://nexustrade.io): AI backtesters. Describe a strategy in plain English, then backtest and paper-trade it. Trust 50 (Preliminary).
- [QuantConnect](https://www.quantconnect.com): AI backtesters. A cloud quant platform on the open-source LEAN engine: backtest, paper and live. Trust 80 (Preliminary).
- [TradersPost](https://traderspost.io): Build your own agent. Turns TradingView or webhook alerts into broker orders. Trust 65 (Preliminary).
- [edgeful](https://www.edgeful.com): ORB / futures optimizers. Opening-range and session statistics for futures and stocks. Trust 0 (Preliminary).
- [TradingAgents](https://github.com/TauricResearch/TradingAgents): Build your own agent. An open-source multi-agent LLM trading framework of analysts, researchers and a trader. Very popular. Trust 65 (Preliminary).
- [Microsoft Qlib](https://github.com/microsoft/qlib): Pine / code libraries. An AI-oriented quant research platform for training and backtesting ML models. Trust 65 (Preliminary).
- [FinRL](https://github.com/AI4Finance-Foundation/FinRL): Pine / code libraries. An open-source deep reinforcement learning framework for trading research. Trust 65 (Preliminary).
- [QuantConnect LEAN](https://github.com/QuantConnect/Lean): Pine / code libraries. An open-source algorithmic trading engine you can self-host. Trust 80 (Preliminary).
- [fxraptor Pine indicators](https://github.com/fxraptor-alpha/pinescript-indicators): Pine / code libraries. Open Pine Script indicators, including a 'fractal model'. Rules are readable. Trust 45 (Preliminary).
- [QuantifiedStrategies](https://www.quantifiedstrategies.com): Signal / education publishers. Publishes many rule-based strategy backtests with full entry and exit rules. Trust 30 (Preliminary).
- [OpenBB](https://github.com/OpenBB-finance/OpenBB): Data / regime filters. An open-source investment research platform and data layer that AI agents can use. Trust 45 (Preliminary).
- [Unusual Breakouts](https://unusualbreakouts.com): Data / regime filters. A macro-regime ('green regime') filter for breakout strategies. Trust 30 (Preliminary).
- [Unusual Whales](https://unusualwhales.com): Data / regime filters. Options flow and market data, also sold inside Robinhood's Agent Apps. Trust 0 (Preliminary).
- [Prospero.ai](https://prospero.ai): Data / regime filters. AI stock signal scores built on institutional-flow data. Trust 0 (Preliminary).
- [TMBB-VBT (reaver1000)](https://github.com/reaver1000/tmbb-vbt): Pine / code libraries. Python vectorbt backtesting and walk-forward engine for NQ/ES futures, with a Consolidation Range breakout port (Pine + Python). The author claims 3 of 4 walk-forward folds pass; that's unverified. Trust 45 (Preliminary).
- [SMC Entry Kit (andriiops)](https://github.com/andriiops/smc-entry-kit): Pine / code libraries. MIT-licensed Pine indicator plus a Python backtester that filters false break-of-structure entries near liquidity (ICT/SMC style), for ES/NQ 5-minute charts. Indicator only, no strategy results claimed. Trust 45 (Preliminary).
- [Pine Script strategies (lepotorico2105)](https://github.com/lepotorico2105/pine-script-strategies): Pine / code libraries. Five MIT-licensed Pine v5 strategies (SMC order block + FVG, Bollinger squeeze, EMA crossover, RSI mean reversion, MACD). Vendor: the repo promotes its own Backtrex backtester. Claimed results (e.g. PF 1.45) exclude costs. Trust 45 (Preliminary).

## US brokers for AI agents (see /brokers.html, /brokers.json)
Automation levels: 1 Full autonomy (official AI-agent path documented to allow unattended orders) > 2 Programmable API (build your own bot; unattended AI agent not officially confirmed) > 3 API automated, AI connector needs approval > 4 Per-order approval required > 5 Manual only / no API. As of Oct 2026, Robinhood is the only broker confirmed hands-on for fully autonomous AI-agent trading; Public and tastytrade are Full autonomy per official docs (not tested by us).
- Robinhood: automation Full autonomy (no human intervention); 0DTE Yes; MCP Official; paper Unverified. Verified 2026-10-17.
- Public: automation Full autonomy (no human intervention); 0DTE Yes; MCP Official; paper Unverified. Verified 2026-10-17.
- tastytrade: automation Full autonomy (no human intervention); 0DTE Yes; MCP Official; paper Yes (API sandbox). Verified 2026-10-17.
- Charles Schwab: automation Programmable API (build your own bot); 0DTE Yes; MCP Community; paper Yes (thinkorswim paperMoney). Verified 2026-10-17.
- Webull: automation Programmable API (build your own bot); 0DTE Yes; MCP Official; paper Yes (API sandbox 'uat'). Verified 2026-10-17.
- E*TRADE (Morgan Stanley): automation Programmable API (build your own bot); 0DTE Unverified; MCP None found; paper Sandbox only (sample data, no simulated fills). Verified 2026-10-17.
- Alpaca: automation Programmable API (build your own bot); 0DTE Yes; MCP Official; paper Yes. Verified 2026-10-17.
- Tradier: automation Programmable API (build your own bot); 0DTE Unverified; MCP Official; paper Yes. Verified 2026-10-17.
- Interactive Brokers (IBKR): automation API automated, AI connector needs approval; 0DTE Unverified; MCP Official; paper Yes. Verified 2026-10-17.
- TradeStation: automation API automated, AI connector needs approval; 0DTE Unverified; MCP Official; paper Yes (SIM API). Verified 2026-10-17.
- Fidelity: automation Manual only / no API; 0DTE Unverified; MCP None found; paper Unverified. Verified 2026-10-17.
- Merrill Edge: automation Manual only / no API; 0DTE Unverified; MCP None found; paper Unverified. Verified 2026-10-17.
- Firstrade: automation Manual only / no API; 0DTE Unverified; MCP None found; paper Unverified. Verified 2026-10-17.

## Strategy prompts (results are original posters' claims; prompts default to paper + per-trade approval)
- [Race AI models on $1k paper portfolios](/prompts/ai-models-paper-portfolio-race): Give several AI models the same daily pre-market prompt and a $1,000 paper account each, then see which one actually trades well.
- [QQQ RSI pullback with trend and VWAP filters](/prompts/qqq-rsi-ema-vwap-pullback): Buy short-term oversold dips in QQQ only while the long-term trend is rising, and sell once momentum recovers.
- [QQQ six-up-days plus low IBS, held 21 days](/prompts/qqq-6-up-days-low-ibs): When QQQ has rallied six days straight but closed near the day's low, buy and hold for about a month.
- [SPY three-down-days overnight bounce](/prompts/spy-3-down-days-overnight): After three straight down closes in SPY, buy at the close and sell at the next morning's open.
- [SPY trim-below-EMAs, rebuy near the 50-day (a cautionary test)](/prompts/spy-ema-trim-50sma-rebuy): Sell SPY when it loses its 8/21-day EMAs and buy back near the 50-day SMA. The poster found it trailed buy-and-hold.
- [QQQ 9/31 EMA trend-follow, stress-tested](/prompts/qqq-9-31-ema-trend): Go long QQQ when price is above a rising fast EMA stack (9 > 31) and short on the mirror signal, with fixed stops and targets.
- [QQQ 5-minute opening range breakout (paper vs. replications)](/prompts/qqq-5-minute-orb): Trade QQQ in the direction of the first 5-minute candle, stop at the other side of that candle, and target 10R or exit at the close.
- [SPX 0DTE iron condor at 10:00 AM](/prompts/spx-0dte-iron-condor-10am): Sell a same-day SPX iron condor with 0.10–0.15 delta short strikes at 10 AM, and stop out when the loss equals the credit received.
- [0DTE iron butterfly with small target and noon exit](/prompts/0dte-iron-butterfly-quick-exit): Open a 0DTE iron butterfly mid-morning, take +15%, cut at −25%, and close by noon.
- [SPY wheel at 45 DTE](/prompts/spy-wheel-45-dte): Sell cash-secured SPY puts about 45 days out; if assigned, sell covered calls until called away, then repeat.
- [52-week-high momentum portfolio with 20-week exit](/prompts/52-week-high-momentum-20w-exit): Buy stocks making fresh 52-week highs a few at a time, and exit any holding whose weekly close drops below its 20-week SMA.
- [NQ 30-minute opening-range breakout with three filters](/prompts/nq-30-min-orb-three-filters): Trade the first 5-minute close outside NQ's 9:30–9:59 ET range, but only on days that pass three context filters.
- [NQ breakout from today's open, sized off yesterday's range](/prompts/nq-open-stop-breakout): Place a buy stop above and a sell stop below NQ's open, spaced by yesterday's range, and be flat by the close.
- [Nasdaq-100 Turtle-style breakouts with a cash regime filter](/prompts/nasdaq-100-turtle-breakout-regime): Buy Nasdaq-100 stocks breaking to new multi-week highs, size by volatility, use a time stop, and go to cash when the market regime turns.
- [QQQ 5-minute VWAP rebound after a weak day](/prompts/qqq-5-min-vwap-rebound): After QQQ has a weak day, buy a small intraday rebound while price is still below the session VWAP, and be out by the close.
- [Stock breakouts vs pullbacks, switched by the SPY 200-day regime](/prompts/breakouts-vs-pullbacks-by-regime): Run tight-base breakouts and confirmed pullbacks on US stocks, and log whether SPY is above or below its 200-day average to see which regime really pays.
- [Strategy reality check: baseline, decay and position-size audit](/prompts/strategy-reality-check-audit): Before trusting any backtest you saw on X, have your agent compare it with buy-and-hold, check whether the edge faded, and stress-test the risk per trade.

## New this week (X scan Oct 11–18, 2026; claims unverified)
- NQ 30-min ORB: three filters turn break-even into PF 2.24 (claimed) (Herman Trading (@RHerman), 2026-10-18): Claims PF 2.31 (2015–21) and 2.24 (2022–26), 230 trades, 62.61% win rate. The author notes that 2022–26 was used to confirm the filters. Source: https://x.com/RHerman/status/2111910235938198004 Prompt: https://agenticalgorank.com/prompts/nq-30-min-orb-three-filters
- NQ stop-order breakout from the open, sized off yesterday's range (AEY Capital (@AEY_Capital), 2026-10-18): Claims 1,747 trades, PF 1.40, 13% max drawdown, and +0.22R/trade out-of-sample since 2022. Source: https://x.com/AEY_Capital/status/2111885279812563193 Prompt: https://agenticalgorank.com/prompts/nq-open-stop-breakout
- Nasdaq-100 Turtle breakouts with a cash regime filter (self-promotional) (SetupAlpha (@SetupAlpha), 2026-10-18): Claims 17.4%/yr 2000–Jul 2026 vs SPY 8.1%, worst drawdown 20.9%. Parameters not public. Source: https://x.com/SetupAlpha/status/2111767608156459204 Prompt: https://agenticalgorank.com/prompts/nasdaq-100-turtle-breakout-regime
- QQQ 5-min VWAP rebound, with costs shown (QuantifiedStrategies (@QuantifiedStrat), 2026-10-14): Claims 298 trades and an 82.6% win rate. PF 2.96 before costs, 2.41 after $0.03/share. Source: https://x.com/QuantifiedStrat/status/2110370038570197260 Prompt: https://agenticalgorank.com/prompts/qqq-5-min-vwap-rebound
- Breakouts paid far more with SPY below its 200-day (Burak Dağkılıç (@burak_dagkilic), 2026-10-17): Claims breakouts averaged 1.23R in a downtrend regime vs 0.16R in an uptrend, from ~1,700 stocks over 2 years with no commissions. Source: https://x.com/burak_dagkilic/status/2111612346187542529 Prompt: https://agenticalgorank.com/prompts/breakouts-vs-pullbacks-by-regime
- Cautionary: 10% risk per trade turned $10k into $38 (BacktestingMax (@BacktestingMax), 2026-10-18): Claims the same 456 RSI trades ended at $11,304 at 1% risk vs $38 at 10%. Source: https://x.com/BacktestingMax/status/2111925559273881867 Prompt: https://agenticalgorank.com/prompts/strategy-reality-check-audit
- Cautionary: a viral VIX buy rule vs just holding SPY (SPXVIX (@SPXVIX), 2026-10-16): Claims $1 became $7.78 with the rule vs $32.09 in SPY since 1993. Source: https://x.com/SPXVIX/status/2111270882845098147 Prompt: https://agenticalgorank.com/prompts/strategy-reality-check-audit
- Cautionary: TLT month-end edge has faded since 2018 (Trading Time Machine (@thetrading), 2026-10-15): Claims Larry Williams' rule earned 0.11%/trade vs 0.10% for any 6-day window, and the last-day edge is down to ~3 bp since 2018. Source: https://x.com/thetrading/status/2110697698269565256 Prompt: https://agenticalgorank.com/prompts/strategy-reality-check-audit
- Larry Connors' Double Seven on SPY (QuantifiedStrategies, 2026-10-18): Rules: long at the close when above the 200-day MA and at a 7-day low, exit on a 7-day high close, no stop. Results are shown only in a chart. Source: https://x.com/QuantifiedStrat/status/2111878728187814294
- TRIN (Arms Index) > 100 buy rule on SPY (QuantifiedStrategies, 2026-10-13): Rules: buy SPY at the close when TRIN > 100, sell when the close is above yesterday's high. Equity curve shown only as an image. Source: https://x.com/QuantifiedStrat/status/2110066788964749777
- Meb Faber 200-day trend filter on the S&P 500 (QuantifiedStrategies, 2026-10-18): Claims CAGR 6.45% vs 7% buy-and-hold, with max drawdown cut to 28%. Source: https://x.com/QuantifiedStrat/status/2111804488994730430
- SPY/EEM/TLT monthly momentum rotation (QuantifiedStrategies, 2026-10-14): Claims it beat buy-and-hold with lower drawdowns until 2022, which was a hard year. Source: https://x.com/QuantifiedStrat/status/2110429176725389634
- Gold bullish engulfing: no edge vs baseline (BacktestingMax, 2026-10-18): Claims gold was up 3 days later 53% of the time after the pattern, the same as on all other days. Source: https://x.com/BacktestingMax/status/2111880012114637169

## Community
- /submit : submit a tool or strategy (human-reviewed queue; nothing auto-published)
- /feed.xml : RSS of new items
- "I want this" votes show community interest only and never change Trust Scores.