--:--:--LOCAL ·--:--UTC
    MKT CAP$2.93T+0.6%
    24H VOL$109.8B
    BTC DOM58.0%
    HEALTH75BULLISH
    SHORT-TERM67NEUTRAL
    LONG-TERM83BULLISH
    OI$14.1B
    24H LIQ$78M
    LONG/SHORT53.9% / 46.1%
    REGIME (LT)BTC-LED BULL MARKET
    REGIME (ST)SQUEEZE
    HL OI$13.0B
    WHALESSHORT 43.8%
    MKT CAP$2.93T+0.6%
    24H VOL$109.8B
    BTC DOM58.0%
    HEALTH75BULLISH
    SHORT-TERM67NEUTRAL
    LONG-TERM83BULLISH
    OI$14.1B
    24H LIQ$78M
    LONG/SHORT53.9% / 46.1%
    REGIME (LT)BTC-LED BULL MARKET
    REGIME (ST)SQUEEZE
    HL OI$13.0B
    WHALESSHORT 43.8%
    Plug real-time crypto data into your AI agent — one command: Get your free API key →
    LoginGet API Key
    Home / Trading Strategies / AI & Machine Learning

    AI & Machine Learning Crypto Trading Strategies

    7 ai & machine learning strategies for crypto, from the AlgoBrain wiki. Models and agents that learn the signal: ML forecasters, LLM agents and AI-native crypto networks. Each one lists the indicators it uses, the Crypto Data API endpoints that feed it and copy-paste prompts for an AI agent to build and backtest it. All of them are in the API: GET /api/v1/strategies?group=ai-machine-learning.

    7 strategies Most-used indicators: Momentum, VIX (CBOE Volatility Index), Open Interest, Funding Rate, Basis

    Every AI & Machine Learning strategy

    AI Agent Token Arbitrage #

    scalp advanced backtest: live structural edgelatency edgeinformational edge

    Trading the AI agent token category — a 2024-2025 crypto sub-sector where autonomous AI agents (powered by GPT-4, Claude, Llama) own crypto wallets, post on social media, and have associated bonding-curve tokens that speculators trade.

    Why it works: AI agent tokens (Virtuals on Base, ai16z on Solana, Truth Terminal-spawned tokens) launch via bonding curves with social-momentum-driven price discovery. Token value is correlated with the agent's social-media virality, which the arb can predict via Twitter/Discord/Farcaster monitoring and triangulate against on-chain bonding curve state.

    Indicators Price and volume only — see the playbook for the exact rules.
    CDA endpoints/api/v1/sentiment/fear-greed/api/v1/dex/trending/api/v1/dex/new-pools/api/v1/hyperliquid/candles
    Via API/api/v1/strategies/ai-agent-token-arbitrage
    AI-agent prompts
    Build it with an AI agent
    Build the AI Agent Token Arbitrage crypto trading strategy using the CryptoDataAPI (X-API-Key header, base https://cryptodataapi.com).
    
    1. Read the playbook first: GET https://cryptodataapi.com/api/v1/algobrain/page?path=wiki/strategies/arbitrage/ai-agent-token-arbitrage.md
    2. Pull the inputs:
    - GET https://cryptodataapi.com/api/v1/sentiment/fear-greed
    - GET https://cryptodataapi.com/api/v1/dex/trending
    - GET https://cryptodataapi.com/api/v1/dex/new-pools
    - GET https://cryptodataapi.com/api/v1/hyperliquid/candles
    3. Compute the signals described in the playbook on 5m bars (pinned: interval=5m, lookback=500 bars, universe=BTC,ETH,SOL unless the playbook says otherwise).
    4. Emit entry/exit rules, position size (risk 1% of equity per trade) and a stop, as JSON: {"symbol","side","entry","stop","target","size_pct","reason"}.
    5. State which regime the rules are valid in (GET /api/v1/regimes/current) and stand aside outside it. Research only — do not place orders.
    Backtest it
    Backtest the AI Agent Token Arbitrage strategy on CryptoDataAPI history before trusting it.
    
    - Bars: GET https://cryptodataapi.com/api/v1/backtesting/klines?symbol=BTC&interval=15m (Pro)
    - Funding: GET https://cryptodataapi.com/api/v1/backtesting/funding?symbol=BTC (Pro)
    - Rules: from the playbook at /api/v1/algobrain/page?path=wiki/strategies/arbitrage/ai-agent-token-arbitrage.md
    Pinned: fees 4.5 bps taker per side, 2 bps slippage, funding applied every 8h, signals on bar close only (no lookahead), 70/30 in-sample/out-of-sample split. Report CAGR, Sharpe, max drawdown, trade count and out-of-sample vs in-sample decay.

    AI Agent Trading Strategies #

    intraday advanced backtest: untested informational edgeanalytical edge

    AI agent strategies use large language models and autonomous agents as components within trading systems. Unlike traditional algorithmic trading where rules are explicitly coded, agent-based strategies leverage LLMs for reasoning, interpretation, and adaptation.

    Why it works: LLM agents parse unstructured text (news, filings, transcripts, social feeds) faster and across more names than discretionary humans, extracting a tradable signal before slower readers reprice the asset.

    CDA endpoints/api/v1/sentiment/fear-greed/api/v1/news/market-moving/api/v1/hyperliquid/candles/api/v1/market-data/klines
    Via API/api/v1/strategies/ai-agent-strategies
    AI-agent prompts
    Build it with an AI agent
    Build the AI Agent Trading Strategies crypto trading strategy using the CryptoDataAPI (X-API-Key header, base https://cryptodataapi.com).
    
    1. Read the playbook first: GET https://cryptodataapi.com/api/v1/algobrain/page?path=wiki/strategies/algorithmic/ai-agent-strategies.md
    2. Pull the inputs:
    - GET https://cryptodataapi.com/api/v1/sentiment/fear-greed
    - GET https://cryptodataapi.com/api/v1/news/market-moving
    - GET https://cryptodataapi.com/api/v1/hyperliquid/candles
    - GET https://cryptodataapi.com/api/v1/market-data/klines
    3. Compute VIX (CBOE Volatility Index), Momentum on 15m bars (pinned: interval=15m, lookback=500 bars, universe=BTC,ETH,SOL unless the playbook says otherwise).
    4. Emit entry/exit rules, position size (risk 1% of equity per trade) and a stop, as JSON: {"symbol","side","entry","stop","target","size_pct","reason"}.
    5. State which regime the rules are valid in (GET /api/v1/regimes/current) and stand aside outside it. Research only — do not place orders.
    Backtest it
    Backtest the AI Agent Trading Strategies strategy on CryptoDataAPI history before trusting it.
    
    - Bars: GET https://cryptodataapi.com/api/v1/backtesting/klines?symbol=BTC&interval=15m (Pro)
    - Funding: GET https://cryptodataapi.com/api/v1/backtesting/funding?symbol=BTC (Pro)
    - Rules: from the playbook at /api/v1/algobrain/page?path=wiki/strategies/algorithmic/ai-agent-strategies.md
    Pinned: fees 4.5 bps taker per side, 2 bps slippage, funding applied every 8h, signals on bar close only (no lookahead), 70/30 in-sample/out-of-sample split. Report CAGR, Sharpe, max drawdown, trade count and out-of-sample vs in-sample decay.

    AI Tokens Basket (Hyperliquid Basket) #

    swing intermediate backtest: untested behavioral edgeinformational edge

    A sector basket of artificial intelligence and AI-agent crypto tokens with active Hyperliquid perpetuals.

    Why it works: AI token prices are driven by narrative velocity around real-world AI breakthroughs (model releases, GPU allocation news, agent deployment milestones); these narrative shocks are observable before fully priced, and the sector co-moves strongly during AI hype cycles — creating momentum entries with defined narrative-reversal exit signals.

    Indicators Open Interest
    CDA endpoints/api/v1/derivatives/funding-rates/api/v1/hyperliquid/funding-rates/api/v1/derivatives/open-interest/api/v1/hyperliquid/open-interest/api/v1/market-health/altcoin-breadth/api/v1/coins/top
    Via API/api/v1/strategies/ai-tokens-basket
    AI-agent prompts
    Build it with an AI agent
    Build the AI Tokens Basket (Hyperliquid Basket) crypto trading strategy using the CryptoDataAPI (X-API-Key header, base https://cryptodataapi.com).
    
    1. Read the playbook first: GET https://cryptodataapi.com/api/v1/algobrain/page?path=wiki/strategies/hyperliquid-baskets/ai-tokens-basket.md
    2. Pull the inputs:
    - GET https://cryptodataapi.com/api/v1/derivatives/funding-rates
    - GET https://cryptodataapi.com/api/v1/hyperliquid/funding-rates
    - GET https://cryptodataapi.com/api/v1/derivatives/open-interest
    - GET https://cryptodataapi.com/api/v1/hyperliquid/open-interest
    - GET https://cryptodataapi.com/api/v1/market-health/altcoin-breadth
    - GET https://cryptodataapi.com/api/v1/coins/top
    3. Compute Open Interest on 4h bars (pinned: interval=4h, lookback=500 bars, universe=BTC,ETH,SOL unless the playbook says otherwise).
    4. Emit entry/exit rules, position size (risk 1% of equity per trade) and a stop, as JSON: {"symbol","side","entry","stop","target","size_pct","reason"}.
    5. State which regime the rules are valid in (GET /api/v1/regimes/current) and stand aside outside it. Research only — do not place orders.
    Backtest it
    Backtest the AI Tokens Basket (Hyperliquid Basket) strategy on CryptoDataAPI history before trusting it.
    
    - Bars: GET https://cryptodataapi.com/api/v1/backtesting/klines?symbol=BTC&interval=4h (Pro)
    - Funding: GET https://cryptodataapi.com/api/v1/backtesting/funding?symbol=BTC (Pro)
    - Rules: from the playbook at /api/v1/algobrain/page?path=wiki/strategies/hyperliquid-baskets/ai-tokens-basket.md
    Pinned: fees 4.5 bps taker per side, 2 bps slippage, funding applied every 8h, signals on bar close only (no lookahead), 70/30 in-sample/out-of-sample split. Report CAGR, Sharpe, max drawdown, trade count and out-of-sample vs in-sample decay.

    AI-Amplified Exploit Arbitrage #

    swing advanced backtest: live structural edgeinformational edgeanalytical edge

    A hub strategy framing for the recurring arbitrage opportunities created by AI-driven smart-contract exploits.

    Why it works: AI-driven vuln discovery is compressing the cost of finding smart-contract exploits while defender deployment latency stays governance-bound. Post-hack market structure is predictable (pause → governance vote → restitution), creating recurring trading windows. Counterparty: forced-sellers (panic LPs, IOU-discount holders) and slow-to-react market-makers.

    Indicators Price and volume only — see the playbook for the exact rules.
    CDA endpoints/api/v1/derivatives/funding-rates/api/v1/hyperliquid/funding-rates/api/v1/on-chain/exchange-flows/spike-alerts/api/v1/news/market-moving/api/v1/dex/trending/api/v1/dex/new-pools
    Via API/api/v1/strategies/ai-amplified-exploit-arbitrage
    AI-agent prompts
    Build it with an AI agent
    Build the AI-Amplified Exploit Arbitrage crypto trading strategy using the CryptoDataAPI (X-API-Key header, base https://cryptodataapi.com).
    
    1. Read the playbook first: GET https://cryptodataapi.com/api/v1/algobrain/page?path=wiki/strategies/arbitrage/ai-amplified-exploit-arbitrage.md
    2. Pull the inputs:
    - GET https://cryptodataapi.com/api/v1/derivatives/funding-rates
    - GET https://cryptodataapi.com/api/v1/hyperliquid/funding-rates
    - GET https://cryptodataapi.com/api/v1/on-chain/exchange-flows/spike-alerts
    - GET https://cryptodataapi.com/api/v1/news/market-moving
    - GET https://cryptodataapi.com/api/v1/dex/trending
    - GET https://cryptodataapi.com/api/v1/dex/new-pools
    3. Compute the signals described in the playbook on 4h bars (pinned: interval=4h, lookback=500 bars, universe=BTC,ETH,SOL unless the playbook says otherwise).
    4. Emit entry/exit rules, position size (risk 1% of equity per trade) and a stop, as JSON: {"symbol","side","entry","stop","target","size_pct","reason"}.
    5. State which regime the rules are valid in (GET /api/v1/regimes/current) and stand aside outside it. Research only — do not place orders.
    Backtest it
    Backtest the AI-Amplified Exploit Arbitrage strategy on CryptoDataAPI history before trusting it.
    
    - Bars: GET https://cryptodataapi.com/api/v1/backtesting/klines?symbol=BTC&interval=4h (Pro)
    - Funding: GET https://cryptodataapi.com/api/v1/backtesting/funding?symbol=BTC (Pro)
    - Rules: from the playbook at /api/v1/algobrain/page?path=wiki/strategies/arbitrage/ai-amplified-exploit-arbitrage.md
    Pinned: fees 4.5 bps taker per side, 2 bps slippage, funding applied every 8h, signals on bar close only (no lookahead), 70/30 in-sample/out-of-sample split. Report CAGR, Sharpe, max drawdown, trade count and out-of-sample vs in-sample decay.

    Alpha Token Arbitrage (Bittensor) #

    intraday advanced backtest: untested structural edgeanalytical edge

    Alpha-token arbitrage is a set of structural trades that exploit pricing inefficiencies in Bittensor's dTAO bonding curves.

    Why it works: Bittensor alpha tokens are priced by subnet-specific bonding curves, while their fundamental value derives from expected future TAO emission share. When the bonding curve lags the emission-share signal, or when the same alpha is listed on multiple third-party venues (Rayon, tao.bit) at different implied prices, riskless or near-riskless arbitrage is available.

    Indicators Price and volume only — see the playbook for the exact rules.
    CDA endpoints/api/v1/hyperliquid/candles/api/v1/market-data/klines
    Via API/api/v1/strategies/alpha-token-arbitrage
    AI-agent prompts
    Build it with an AI agent
    Build the Alpha Token Arbitrage (Bittensor) crypto trading strategy using the CryptoDataAPI (X-API-Key header, base https://cryptodataapi.com).
    
    1. Read the playbook first: GET https://cryptodataapi.com/api/v1/algobrain/page?path=wiki/strategies/alpha-token-arbitrage.md
    2. Pull the inputs:
    - GET https://cryptodataapi.com/api/v1/hyperliquid/candles
    - GET https://cryptodataapi.com/api/v1/market-data/klines
    3. Compute the signals described in the playbook on 15m bars (pinned: interval=15m, lookback=500 bars, universe=BTC,ETH,SOL unless the playbook says otherwise).
    4. Emit entry/exit rules, position size (risk 1% of equity per trade) and a stop, as JSON: {"symbol","side","entry","stop","target","size_pct","reason"}.
    5. State which regime the rules are valid in (GET /api/v1/regimes/current) and stand aside outside it. Research only — do not place orders.
    Backtest it
    Backtest the Alpha Token Arbitrage (Bittensor) strategy on CryptoDataAPI history before trusting it.
    
    - Bars: GET https://cryptodataapi.com/api/v1/backtesting/klines?symbol=BTC&interval=15m (Pro)
    - Funding: GET https://cryptodataapi.com/api/v1/backtesting/funding?symbol=BTC (Pro)
    - Rules: from the playbook at /api/v1/algobrain/page?path=wiki/strategies/alpha-token-arbitrage.md
    Pinned: fees 4.5 bps taker per side, 2 bps slippage, funding applied every 8h, signals on bar close only (no lookahead), 70/30 in-sample/out-of-sample split. Report CAGR, Sharpe, max drawdown, trade count and out-of-sample vs in-sample decay.

    Bittensor Subnet Rotation #

    swing advanced backtest: untested informational edgebehavioral edge

    Subnet rotation is a Bittensor-specific strategy that reallocates capital across subnet alpha tokens in response to changes in each subnet's block-to-block emission share. Post-dtao (Feb 2025) emission share is market-determined via alpha bonding curves, which makes it a high-frequency observable.

    Why it works: Subnet emission share on Bittensor reallocates block-by-block via dTAO bonding curves; most market participants cannot price the subnet-fundamentals signal fast enough, so rotation between alpha tokens captures the share-gain before the bonding curve prices it in.

    Indicators Momentum
    CDA endpoints/api/v1/hyperliquid/candles/api/v1/market-data/klines
    Via API/api/v1/strategies/bittensor-subnet-rotation
    AI-agent prompts
    Build it with an AI agent
    Build the Bittensor Subnet Rotation crypto trading strategy using the CryptoDataAPI (X-API-Key header, base https://cryptodataapi.com).
    
    1. Read the playbook first: GET https://cryptodataapi.com/api/v1/algobrain/page?path=wiki/strategies/bittensor-subnet-rotation.md
    2. Pull the inputs:
    - GET https://cryptodataapi.com/api/v1/hyperliquid/candles
    - GET https://cryptodataapi.com/api/v1/market-data/klines
    3. Compute Momentum on 4h bars (pinned: interval=4h, lookback=500 bars, universe=BTC,ETH,SOL unless the playbook says otherwise).
    4. Emit entry/exit rules, position size (risk 1% of equity per trade) and a stop, as JSON: {"symbol","side","entry","stop","target","size_pct","reason"}.
    5. State which regime the rules are valid in (GET /api/v1/regimes/current) and stand aside outside it. Research only — do not place orders.
    Backtest it
    Backtest the Bittensor Subnet Rotation strategy on CryptoDataAPI history before trusting it.
    
    - Bars: GET https://cryptodataapi.com/api/v1/backtesting/klines?symbol=BTC&interval=4h (Pro)
    - Funding: GET https://cryptodataapi.com/api/v1/backtesting/funding?symbol=BTC (Pro)
    - Rules: from the playbook at /api/v1/algobrain/page?path=wiki/strategies/bittensor-subnet-rotation.md
    Pinned: fees 4.5 bps taker per side, 2 bps slippage, funding applied every 8h, signals on bar close only (no lookahead), 70/30 in-sample/out-of-sample split. Report CAGR, Sharpe, max drawdown, trade count and out-of-sample vs in-sample decay.

    Crypto Shorts in an AI-Driven Recession #

    position advanced backtest: untested behavioral edgestructural edgeinformational edge

    A basket of crypto-specific short and pair-trade legs designed to express the ai-recession-playbook thesis through digital assets.

    Why it works: Crypto markets price BTC and alts as a single risk-on asset class, but AI labor recession produces dispersion: tech-worker wealth destruction and VC dry-up hit alts/AI-tokens/mining-equities asymmetrically while BTC may decouple on Fed easing. The trade is the dispersion, not direction.

    CDA endpoints/api/v1/derivatives/funding-rates/api/v1/hyperliquid/funding-rates/api/v1/market-health/altcoin-breadth/api/v1/coins/top/api/v1/market-intelligence/etf/{asset}/flows/api/v1/sentiment/macro
    Via API/api/v1/strategies/crypto-ai-recession-shorts
    AI-agent prompts
    Build it with an AI agent
    Build the Crypto Shorts in an AI-Driven Recession crypto trading strategy using the CryptoDataAPI (X-API-Key header, base https://cryptodataapi.com).
    
    1. Read the playbook first: GET https://cryptodataapi.com/api/v1/algobrain/page?path=wiki/strategies/crypto-ai-recession-shorts.md
    2. Pull the inputs:
    - GET https://cryptodataapi.com/api/v1/derivatives/funding-rates
    - GET https://cryptodataapi.com/api/v1/hyperliquid/funding-rates
    - GET https://cryptodataapi.com/api/v1/market-health/altcoin-breadth
    - GET https://cryptodataapi.com/api/v1/coins/top
    - GET https://cryptodataapi.com/api/v1/market-intelligence/etf/{asset}/flows
    - GET https://cryptodataapi.com/api/v1/sentiment/macro
    3. Compute Funding Rate, Basis, Liquidation on 1d bars (pinned: interval=1d, lookback=500 bars, universe=BTC,ETH,SOL unless the playbook says otherwise).
    4. Emit entry/exit rules, position size (risk 1% of equity per trade) and a stop, as JSON: {"symbol","side","entry","stop","target","size_pct","reason"}.
    5. State which regime the rules are valid in (GET /api/v1/regimes/current) and stand aside outside it. Research only — do not place orders.
    Backtest it
    Backtest the Crypto Shorts in an AI-Driven Recession strategy on CryptoDataAPI history before trusting it.
    
    - Bars: GET https://cryptodataapi.com/api/v1/backtesting/klines?symbol=BTC&interval=1d (Pro)
    - Funding: GET https://cryptodataapi.com/api/v1/backtesting/funding?symbol=BTC (Pro)
    - Rules: from the playbook at /api/v1/algobrain/page?path=wiki/strategies/crypto-ai-recession-shorts.md
    Pinned: fees 4.5 bps taker per side, 2 bps slippage, funding applied every 8h, signals on bar close only (no lookahead), 70/30 in-sample/out-of-sample split. Report CAGR, Sharpe, max drawdown, trade count and out-of-sample vs in-sample decay.

    Get these strategies from the API

    curl -H "X-API-Key: cdk_live_yourkey" \
      "https://cryptodataapi.com/api/v1/strategies?group=ai-machine-learning"
    
    curl -H "X-API-Key: cdk_live_yourkey" \
      "https://cryptodataapi.com/api/v1/strategies/ai-agent-token-arbitrage"

    Any key works, Free included — mint one in a single call. The list endpoint returns summaries; the per-slug endpoint adds the prompts, edge mechanism and data inputs. Full playbooks come from /api/v1/algobrain/page?path=… using each entry's wiki_path. Or use the MCP server.

    What are ai & machine learning crypto trading strategies?

    Models and agents that learn the signal: ML forecasters, LLM agents and AI-native crypto networks.

    How many ai & machine learning strategies are there?

    7: AI Agent Token Arbitrage, AI Agent Trading Strategies, AI Tokens Basket (Hyperliquid Basket), AI-Amplified Exploit Arbitrage, Alpha Token Arbitrage (Bittensor), Bittensor Subnet Rotation, Crypto Shorts in an AI-Driven Recession.

    Which indicators do ai & machine learning strategies use?

    Most often Momentum, VIX (CBOE Volatility Index), Open Interest, Funding Rate.

    Can an AI agent build these strategies from an API?

    Yes. GET /api/v1/strategies?group=ai-machine-learning lists them; GET /api/v1/strategies/{slug} returns the build and backtest prompts, and each prompt names the exact Crypto Data API endpoints to call. Any API key works, Free included.