AI Fund — Open-Source AI Hedge Fund for Claude Code

Hire your AI trading desk. Fire the ones that miss KPIs.

45 AI trading agents. 21 named personas (Arthur Hayes, Jim Simons, George Soros, Jesse Livermore, Warren Buffett, Peter Lynch…). 147 built-in MCP tools across Cube (31), CCXT (92), and Alpaca (24), plus 110 CCXT exchanges. Paper trading by default. MIT licensed. Runs on Claude Code.

> /hire risk-manager
> /hire arthur-hayes
> /hire market-maker

> @arthur-hayes what's the macro thesis right now?
> scan all exchanges for BTC price differences
> the arbitrageur found a 15bps spread between cube.exchange and Binance — execute it
> risk-manager, approve this trade

What Is ai-fund?

45 autonomous trading agents inside Claude Code, with built-in MCP connectors and broad multi-exchange support via CCXT. Shared analysis libraries cover indicators, execution, portfolio analytics, and risk tooling.

21 are named personas — Arthur Hayes, Jim Simons, George Soros, Jesse Livermore, Stanley Druckenmiller, Warren Buffett. The other 24 are role-based: scalpers, market makers, risk managers, quants, arbitrageurs.

No config files. No YAML. You hire agents that fit your thesis and fire the ones that don't deliver. Each one carries its own personality, philosophy, and KPIs.

Count Snapshot (from this repo)

45 agents total in skills/ (_template excluded)
21 named personas + 24 role-based specialists
147 built-in MCP tools across active servers
- Cube: 31 tools
- CCXT: 92 tools
- Alpaca: 24 tools
110 exchanges via CCXT (from the installed ccxt package in this repo)

How is this different from a grid bot?

You get a quant analyst that only trusts data, a risk manager that blocks trades when the sizing is wrong, and a market maker running Avellaneda-Stoikov across three venues.

The arbitrageur watches every connected exchange for mispricings and won't shut up about it. They argue with each other. The risk manager says no a lot.

More exchanges = smarter desk. Cross-exchange arb, smart order routing, multi-venue MM.

What You Can Do

Strategy	Description
Cross-exchange arb	Spot price gaps, execute both legs
Market making	Multi-venue quotes, Avellaneda-Stoikov
Macro trading	DXY, yields, Fed policy via Hayes agent
Stat arb / quant	Mean reversion, momentum, pairs
Portfolio	Risk parity, Kelly sizing, drawdowns
Execution algos	TWAP, VWAP, Iceberg routing

How It Works

Step	What Happens
1. Connect	Connect any exchange via plugins. Paper mode by default.
2. Hire	Pick agents. KPIs get tracked.
3. Trade	Agents propose, risk manager approves.

Paper trading is on by default. You have to opt in to live.

YOU (trader)
  │
  ├── /hire risk-manager          ← activate agents
  ├── /hire arbitrageur
  ├── /hire market-maker
  │
  ▼
CLAUDE CODE (AI runtime)
  │
  ├── Skills (45 SKILL.md files)  ← agent personas, strategies, KPIs
  │
  ├── Exchange Connectors (MCP)   ← connect any exchange
  │   ├── Cube (built-in)
  │   ├── Binance, Coinbase, Kraken, OKX...
  │   └── 110 via CCXT
  │
  ▼
YOUR EXCHANGES (paper or live)

Skills define what an agent thinks and does.

Connectors talk to exchanges.

The two layers don't know about each other. Add an exchange, don't touch agent code. Write an agent, don't touch exchange code.

Who Is This For?

You are a...	ai-fund gives you...
Crypto trader	45 agents, natural language
Quant	Backtest, stat tools, multi-exchange
Fund operator	KPIs, hire/fire, risk controls
Developer	MIT skill system, any exchange

Quick Start

git clone https://github.com/cubexch/ai-fund
cd ai-fund
npm ci

Open Claude Code and connect your exchanges:

claude
> /setup

Hire your first agents:

> /hire risk-manager
> /hire arthur-hayes
> /hire jim-simons

Put them to work:

> @arthur-hayes what's the macro thesis? DXY is falling and the Fed paused.
> @jim-simons scan for statistical anomalies across BTC pairs on all exchanges
> @risk-manager size a long position given current portfolio

Cube CLI + Expanded Tool Surface

Recent Cube connector updates added a first-party cube CLI and expanded analysis/trade tooling.

cd connectors/cube/mcp-server
npm run login
npm run status
npm run cube -- --help

Use the command groups directly (cube account ..., cube market ..., cube order ..., cube risk ..., cube trade ...) or call MCP tools from Claude Code. Full reference: connectors/cube/README.md.

Advanced Tool Families (Cube)

Portfolio/Risk intelligence: assess_portfolio_risk, simulate_stress_test
Signal intelligence: detect_confluence, detect_bb_squeeze, get_market_microstructure
Execution planning: plan_twap, simulate_market_impact
Smart routing/discovery: search_assets, get_trending, execute_trade

These tools are designed to be chained (scan → validate → risk-check → execution plan → route).

Supported Exchanges (Counted)

Connect any exchange via MCP plugins. Each connector handles authentication differently — some use API keys, others use local auth. See connectors/README.md for setup details.

When handing an AI agent access to exchange credentials, think about security. See the API Key Security section below.

Crypto Exchanges

	Cube	Coinbase	Binance	Kraken	OKX
Connector	Built-in MCP	Built-in (CCXT)	Built-in (CCXT)	CLI	okx-mcp
Auth	Local (no keys)	Key+secret+pass	Key+secret	Key+secret	Key+secret+pass
Spot	✅	✅	✅	✅	✅
Perps	✅	Limited	✅	✅	✅
Paper	✅	Sandbox	Testnet	✅	Demo

Equities and Multi-Asset Platforms

	Alpaca	Kraken	IBKR
Setup	Built-in MCP	CLI	CCXT/custom
API keys	Key+secret	Key+secret	Portal auth
Stocks	✅ No-fee	Tokenized	✅
Crypto	✅	✅	❌
Options	❌	❌	✅
Paper	✅	✅	✅

API Key Security — Why This Matters With AI Agents

AI agents can read files, call tools, log output, and spawn processes. Your API key in a config file? The agent can see it.

Risk	Mitigation
Keys in config	Use read-only keys. Disable withdrawal.
Keys in env vars	Use subaccounts with limited funds.
Keys in logs	Avoid verbose mode. Review MCP code.
Keys in transcripts	Scrub before sharing sessions.
No rotation	Rotate regularly. IP whitelist.
Withdrawal enabled	Always disable. Use subaccounts.

Some connectors (like Cube's built-in MCP) use local auth with no API keys. Others require key+secret in config files. Choose connectors that match your security requirements.

110 additional exchanges work via npm i -g ccxt-mcp — anything CCXT supports in this repo's pinned CCXT version.

More venues = more strategies. Cross-exchange arb doesn't work with one exchange.

See connectors/README.md for setup details.

45 AI Trading Agents — The Full Roster

Named Personas

21 agents modeled after real traders. The philosophy isn't just flavor text — it changes how they read markets and size positions.

Persona	Philosophy	Style
Arthur Hayes	Macro-to-crypto. DXY, real yields, liquidity cycles.	Leveraged macro conviction
George Soros	Reflexivity theory. Attack regime breaks. Boom-bust cycles.	Thesis-driven, concentrated
Stanley Druckenmiller	Go for the jugular. Concentrated macro bets when conviction is high.	High-conviction sizing
Paul Tudor Jones	Risk management IS the strategy. 200-day MA. 5:1 R:R minimum.	Trend following, risk-first
Ray Dalio	All-weather portfolio. Risk parity. 15 uncorrelated bets.	Balanced allocation
Jim Simons	Pure quant. Statistical edge. Zero emotion. Sharpe > 2.0.	Systematic stat arb
Ed Thorp	Kelly criterion. Mathematical edge. The original quant.	Optimal bet sizing
Jesse Livermore	Tape reading. Pyramiding. "It was my sitting that made the big money."	Classic speculation
Michael Saylor	Bitcoin is digital property. Stack sats. Never sell.	Relentless BTC accumulation
Cathie Wood	Disruptive innovation. Wright's law. 5-year thesis.	High-conviction innovation
Raoul Pal	Exponential age. Network value. 4-year cycles.	Cycle-based portfolio
PlanB	Stock-to-Flow. Halving cycles. On-chain models.	Model-based BTC valuation
Willy Woo	On-chain analytics. NVT. Holder behavior. "The chain doesn't lie."	On-chain signals
CZ	Build in the bear. Spot only. Ecosystem investing. Fundamentals > hype.	Ecosystem value investing
GCR	Contrarian. Fade the crowd. "When everyone agrees, they're usually wrong."	Contrarian conviction
Cobie	Narrative trading. Early to the meta. Asymmetric bets.	Narrative lifecycle
Ansem	Early discovery. Momentum alpha. Degen with discipline.	Micro-cap momentum
Hsaka	Chart structure. S/R levels. Only A+ setups. Patience.	Technical swing trading
Tetranode	DeFi yield. Real yield vs emissions. Governance power.	Yield optimization
Warren Buffett	Value investing. Margin of safety. Competitive moats. Long-term compounding.	Fundamental value
Gwyneth Chen	Pro market maker. Spread capture. Adverse selection. Avellaneda-Stoikov.	Institutional MM

> /hire arthur-hayes
> @arthur-hayes what's the macro setup? DXY is falling and the Fed just paused.

> /hire jim-simons
> @jim-simons scan for statistical anomalies across all BTC pairs

> /hire michael-saylor
> @michael-saylor set up a weekly DCA into BTC across all exchanges

Role-Based Agents

23 agents organized by function. These don't have celebrity personas — they just do their job.

Active Traders

Agent	Role	Multi-Exchange
Scalper	Sub-second, order book	Lowest-latency venue
Momentum	Breakouts, trend riding	Cross-venue scans
Mean Reversion	Fades extremes	Cross-venue deviation
Swing	Multi-day S/R holds	Best fill routing
Arbitrageur	Buy low, sell high	Core cross-exchange
Grid	Systematic levels	Grid per venue

Execution

Agent	Role	Multi-Exchange
Execution Trader	TWAP, VWAP, Iceberg	Smart order routing
Market Maker	Two-sided quotes	Multi-venue quoting
DCA Strategist	Scheduled buys	Cheapest venue

Research

Agent	Role	Multi-Exchange
Quant Analyst	RSI, MACD, backtests	Cross-venue signals
Order Flow	Tape reading, whales	Cross-venue flow
Volatility	Vol regime detection	Cross-venue vol
Sentiment	Funding, OI, fear/greed	Aggregated data
On-Chain	Wallets, exchange flows	Exchange-agnostic

Risk and Portfolio

Agent	Role	Multi-Exchange
Risk Manager	VaR, Kelly, drawdown caps	Aggregate all venues
Equity Risk Manager	Equity-specific risk: beta, sector, factor	Cross-exchange
Portfolio Manager	Allocation, rebalancing	Cross-exchange
Performance	Post-trade analysis	Per-venue comparison

Specialists

Agent	Role	Multi-Exchange
Funding Farmer	Delta-neutral yield	Best rates cross-venue
Liquidation Hunter	Margin monitoring	All exchanges
Pairs Trader	Long/short correlated	Cross-exchange pairs
Breakout	Range breaks + volume	Cross-venue volume

Infrastructure

Agent	Role	Multi-Exchange
Backtester	Historical simulation	Any exchange data

Performance Evaluation — Hire and Fire Agents Based on KPIs

Every agent has KPIs. Miss them and you're out. /review runs a desk-wide evaluation:

> /review

╔═══════════════════════════════════════════════════╗
║              DESK PERFORMANCE REVIEW              ║
╠═══════════════════════════════════════════════════╣

  CONNECTED: cube.exchange (live) · Binance (paper) · Kraken (paper)

┌──────────────────┬────────────┬────────┬──────────┐
│ Agent            │ Primary KPI│ Actual │ Grade    │
├──────────────────┼────────────┼────────┼──────────┤
│ Risk Manager     │ Breaches   │ 0      │ 🟢 A     │
│ Arbitrageur      │ Net P&L    │ +$340  │ 🟢 B+    │
│ Market Maker     │ Spread P&L │ +$120  │ 🟢 B     │
│ Momentum Trader  │ Win Rate   │ 41%    │ 🔴 D     │
└──────────────────┴────────────┴────────┴──────────┘

RECOMMENDATION:
  🔴 FIRE Momentum Trader — win rate below 55% target
     Market is range-bound. Replace with Mean Reversion Trader.

What ships with each agent:

Component	Description
Metrics	Win rate, Sharpe, spread, drawdown
Self-Eval	Agent grades its own session
Fire Triggers	Hard thresholds, auto-removal

ai-fund vs Other AI Trading Bots

	ai-fund	ai-hedge-fund	Freqtrade	Hummingbot
LLM-native	✅ Claude	✅ Multi-LLM	❌	❌
Agents	45	18	User-defined	~12
Hire/fire	✅	❌	❌	❌
Personas	21	✅	❌	❌
Exchanges	110+	Stocks only	30+	20+
Cross-arb	✅	❌	❌	❌
SOR	✅	❌	❌	❌
Crypto	✅	❌	✅	✅
Multi-MM	✅	❌	❌	1 venue
No API keys	✅ cube	❌	❌	❌
Paper	✅ All	❌	✅	✅
License	MIT	MIT	GPL	Apache

Commands

Command	Description
`/setup`	Connect exchanges, API keys, mode
`/desk`	Active agents, positions, KPIs
`/hire <role>`	Activate an agent
`/fire <role>`	Remove an agent
`/review`	Performance review + fire recs
`/backtest`	Test on historical data

Example Desk Configurations

Desk	Agents
Conservative	risk-manager, dca, performance
Arb	risk, arbitrageur, execution, quant
MM	risk, market-maker, orderflow, vol
Macro	hayes, pal, risk, execution
BTC Maxi	saylor, plan-b, willy-woo, risk
Full	risk, portfolio, arb, mm, hayes, simons

Architecture

ai-fund/
├── connectors/              # Exchange connections (built-in + beta, 110 via CCXT)
│   ├── cube/                # Built-in: cube.exchange (CLI + MCP tool surface)
│   ├── ccxt/                # Built-in: Coinbase, Binance, 110 exchanges (92 tools)
│   ├── alpaca/              # Built-in: stocks, ETFs, crypto (24 tools)
│   └── README.md            # How to add more exchanges
├── skills/                  # 45 agent personas (exchange-agnostic)
├── lib/                     # 28 shared libraries, 250+ pure functions
│   ├── indicators.ts        # SMA, EMA, RSI, MACD, BB, ATR, ADX, OBV, Stochastic
│   ├── math.ts              # Kelly, VaR, Sharpe, Sortino, correlation, drawdown
│   ├── execution-planner.ts # TWAP, VWAP, iceberg, market impact (Almgren-Chriss)
│   ├── execution-analytics.ts # Order book analysis, slippage, fill simulation
│   ├── portfolio-analytics.ts # Portfolio exposure, stress test, rebalancing
│   ├── confluence-detector.ts # Multi-TF confluence, BB squeeze, mean reversion
│   ├── grid-trading.ts      # DCA scheduling, grid optimization, basis trade
│   ├── volume-profile.ts    # Volume profile, value area, correlation regime
│   └── ...                  # Backtester, regime detector, signal generator, etc.
├── examples/                # Pre-built desk configurations
├── scripts/                 # npx installer
└── .claude/commands/        # Slash commands (/setup, /desk, /hire, etc.)

Layer	Role
`skills/`	45 agent personalities, strategies, KPIs
`connectors/`	Built-in and beta exchange MCP servers with capability-gated surfaces
`lib/`	28 shared libraries — indicators, risk, execution, portfolio, options, stat-arb, microstructure
`.claude/commands/`	Slash commands

Add an exchange — no agent files change. Write an agent — no exchange code involved.

FAQ

What is ai-fund?

An open-source AI crypto trading framework with 45 agents running inside Claude Code. You hire the ones that match your strategy and fire the ones that miss KPIs. Think of it as a trading desk, not a bot.

How many trading agents does ai-fund have?

21 named personas (Arthur Hayes, Jim Simons, George Soros, Jesse Livermore, Warren Buffett, and more) plus role-based agents across desk functions. They share analysis libraries for indicators, risk, execution, portfolio analytics, market microstructure, and stat-arb workflows.

What exchanges work with ai-fund?

110 exchanges via CCXT in the pinned dependency, plus dedicated connectors for Cube, Alpaca, Robinhood, Hyperliquid, and gateway orchestration.

Is ai-fund free?

MIT-licensed, fully open source. You need Claude Pro or Team ($20/month) for the Claude Code runtime.

Does ai-fund support multi-exchange trading?

Yes. The Arbitrageur scans for price gaps. The Execution Trader routes to the best venue. The Market Maker quotes across venues at once. It's one of the main reasons to use this.

How is ai-fund different from virattt's ai-hedge-fund?

virattt's project does stocks with investor personas (Buffett, etc.). ai-fund is crypto, works with any exchange, has 45 agents with connector-backed tool surfaces, and fires them when they underperform. Comparison table.

Can ai-fund trade live?

Yes. Everything starts in paper/testnet. The Risk Manager reviews all trades. You have to explicitly confirm before anything goes live.

Does ai-fund work for stocks?

If the exchange supports them. Kraken has tokenized stocks. Alpaca does US equities with full paper trading support.

Which exchange should I start with?

Any exchange you already use. The architecture is venue-agnostic — agents work the same way regardless of connector. If you want zero-config to try things out, Cube's built-in connector needs no API keys. For maximum exchange coverage, use CCXT.

Are my API keys safe with AI agents?

It depends on the connector. Some use local auth (no keys in files). Others require API keys in config — AI agents can read those. Use read-only keys, disable withdrawals, scope to subaccounts, and don't share session transcripts without scrubbing. Full breakdown here.

How do I add my own agent?

Drop a folder in skills/ with a SKILL.md file. Template at skills/_template/SKILL.md.

Building Your Own Agent

Create a folder in skills/ with a SKILL.md file. Use skills/_template/SKILL.md to start.

Section	What Goes In It
Personality	Who they are, how they talk
Philosophy	Beliefs that drive decisions
Capabilities	Which tools, how used
Metrics	KPIs, red flags, fire triggers
Self-Eval	How they grade themselves

Development

Package manager: use npm workspaces only. The governed root commands assume npm ci and npm run ... from the repo root.

CI/CD: GitHub Actions runs npm run build and npm run check on every push and PR (.github/workflows/test.yml). The default PR path covers governed workspaces and hermetic suites only.

Canonical commands:

npm run build builds the shared lib plus supported connector workspaces.
npm run check runs workspace typecheck, hermetic unit tests, hermetic integration tests, and evals.
npm run test:unit runs deterministic default suites.
npm run test:integration runs mocked or replay-backed connector integration suites.
npm run test:platform runs loopback/keychain/process-sensitive suites and is opt-in.
npm run test:live runs real exchange tests and is opt-in.

Auth: Agents authenticate via Device Authorization (RFC 8628) — no API keys needed. See docs/agent-auth-brief.md.

Contributing

See CONTRIBUTING.md. New agents, exchange connectors, bug fixes all welcome.

Disclaimer

ai-fund is for educational and research purposes. Not financial advice. Crypto trading can lose you money. Use paper mode when testing. Backtests don't predict anything.

License

MIT.

Name		Name	Last commit message	Last commit date
Latest commit History 128 Commits
.claude/commands		.claude/commands
.github		.github
bin		bin
connectors		connectors
docs		docs
evals		evals
examples		examples
lib		lib
scripts		scripts
skills		skills
templates		templates
.gitignore		.gitignore
CLAUDE.md		CLAUDE.md
CONTRIBUTING.md		CONTRIBUTING.md
README.md		README.md
SECURITY.md		SECURITY.md
package-lock.json		package-lock.json
package.json		package.json
tsconfig.json		tsconfig.json

Folders and files

Latest commit

History

Repository files navigation

AI Fund — Open-Source AI Hedge Fund for Claude Code

Hire your AI trading desk. Fire the ones that miss KPIs.

What Is ai-fund?

Count Snapshot (from this repo)

How is this different from a grid bot?

What You Can Do

How It Works

Who Is This For?

Quick Start

Cube CLI + Expanded Tool Surface

Advanced Tool Families (Cube)

Supported Exchanges (Counted)

Crypto Exchanges

Equities and Multi-Asset Platforms

API Key Security — Why This Matters With AI Agents

45 AI Trading Agents — The Full Roster

Named Personas

Role-Based Agents

Active Traders

Execution

Research

Risk and Portfolio

Specialists

Infrastructure

Performance Evaluation — Hire and Fire Agents Based on KPIs

ai-fund vs Other AI Trading Bots

Commands

Example Desk Configurations

Architecture

FAQ

What is ai-fund?

How many trading agents does ai-fund have?

What exchanges work with ai-fund?

Is ai-fund free?

Does ai-fund support multi-exchange trading?

How is ai-fund different from virattt's ai-hedge-fund?

Can ai-fund trade live?

Does ai-fund work for stocks?

Which exchange should I start with?

Are my API keys safe with AI agents?

How do I add my own agent?

Building Your Own Agent

Development

Contributing

Disclaimer

License

Links

About

Topics

Resources

Contributing

Security policy

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

Packages