What Is an AI Hedge Fund? A Plain Guide to AI Funds and Agent Teams
An AI hedge fund is an investment setup where AI models help choose and size trades. The term covers professional funds that use machine learning, open-source agent teams like ai-hedge-fund and TradingAgents that simulate a fund for research, and personal setups where one person lets an AI agent trade a small budget under strict limits.
What "AI hedge fund" means (three kinds)
Searchers mix three different things under ai hedge fund, hedge fund ai, and ai hedge funds. A professional fund that uses AI is run by licensed managers for outside investors, with real money and a regulator in the picture. An open-source agent team is a codebase people study. A personal AI fund is one person, their own money, small and capped. A personal setup is not a hedge fund in the legal sense. It is your own account run with AI help. This page does not repeat holdings or return claims about any named fund.
| Kind | Who runs it | Real money? | Example |
|---|---|---|---|
| Professional fund that uses AI | Licensed managers for outside investors | Yes, regulated | A fund you read about in the news. Check that article, not this page, for what it reported. |
| Open-source agent team | Developers and learners | No: ai-hedge-fund does not make trades; TradingAgents sends orders to a simulated exchange | virattt/ai-hedge-fund, TradingAgents |
| Personal AI fund | One person, their own money | Yes, small and capped | Research agents plus one execution agent with limits |
How open-source AI hedge fund projects work
The searches ai hedge fund github, tradingagents, tradingagents github and trading agents github land on two repos. Both are research tools. Neither places a live order on a brokerage or a chain.
virattt/ai-hedge-fund
Per its README, this is a "proof of concept for an AI-powered hedge fund," "for educational purposes only and is not intended for real trading or investment," and "the system does not actually make any trades." Investor-style agents analyze tickers. You need a model API key (Anthropic, OpenAI, DeepSeek, Google, xAI, Kimi and others) and a Financial Datasets key. Install with pipx install aihf. It has backtesting and paper trading modes. GitHub showed 64,453 stars, the MIT license, and a last push on November 2, 2026, when checked on November 3, 2026. The listing on this site is ai-hedge-fund. Qlib and backtesting.py are other research tools people put next to a paper run.
TradingAgents
Per its README, TradingAgents is a multi-agent framework that "mirrors the dynamics of real-world trading firms." The analyst team covers fundamentals, sentiment, news and technicals. Bull and bear researchers debate. A trader proposes. A risk management team and a portfolio manager approve or reject. "If approved, the order will be sent to the simulated exchange and executed." The project is "designed for research purposes." GitHub showed 120,156 stars, the Apache-2.0 license, v0.7.1 (October 2026), and a last push on October 27, 2026. The listing is TradingAgents.
What these projects teach, and what they do not
They show how to split research into roles and make an agent explain its reasoning. They do not prove a strategy makes money. Backtests can overfit, and model output varies run to run. TradingAgents' README says performance "may vary based on many factors." An ai portfolio manager or ai fund manager title on a repo is a role name inside the simulation, not a license to manage someone else's money.
How people run a personal AI fund onchain
These steps apply to any tool, including ones that are not Moby.
- Write a mandate: what you trade, max position size, daily budget, when to exit.
- Do research with agents (open-source teams, your own prompts, smart-money data).
- Use one execution agent with its own wallet and a budget you can afford to lose. Keep your main holdings elsewhere.
- Enforce limits outside the model (server-side), not just in the prompt. See AI agent guardrails.
- Set exits on every position. See AI trading risk management.
- Log every trade and review weekly. Let AI trade for you is the longer version of handing an order to an agent.
Example: a personal AI fund with a Moby Agent
Moby Agents trade Solana tokens, paid in USDC, including tokenized stocks. Examples from the liquid list: SPYx, QQQx, NVDAx and GLDx. See tokenized stocks. Each user gets one agent, with its own wallet separate from your main in-app wallet. The agent cannot withdraw or send funds. To "run a fund" you keep research in other tools and let the one Moby Agent execute. You set your agent's limits on max per trade, daily buy volume and market-cap range from the Moby app, and Moby checks them server-side on every buy your agent makes. That is step 4. Tell your agent your exits and it places real take-profit and stop-loss orders on Moby, including a one-call bracket. Agent buys have no automatic exits, so tell the agent your exits. If you would rather follow a trader with exits attached automatically, Moby copy trading attaches take-profit and stop-loss automatically. See copy trading.
Connect it from Codex, Claude Code (CLI), Gemini CLI (with a paid Gemini API key or enterprise access) and Grok CLI, with the agent key. In ChatGPT, paste your personal Moby connector URL from the app. ChatGPT: works on Business, Enterprise and Edu workspaces that allow custom MCP servers (web). A Moby Agent is a side agentic wallet next to your main trading app, not a replacement. Set up a Moby Agent with its own wallet and your spending guardrails, then connect it to the AI you use.
A mandate you can actually check is short. Name the tokens, a max buy in USDC, a daily buy cap, a market-cap floor and a market-cap ceiling, and the exit you want on each position. If the research notes and the execution agent disagree, the written mandate wins. Review the log once a week and revoke the key if the log does not match the mandate.
A news story about a fund is about that fund's managers and their investors. A GitHub repo is a way to see how an agent team is wired. Neither one is a reason to size a personal account as if a backtest were a promise. Size the personal account so a full loss is an amount you already decided you can lose. An ai portfolio manager inside a simulator is a role in the code. It is not a person you hired.
Risks and what to watch
Model errors and made-up data. Prompt injection from web content. Overfitted backtests. Fees adding up from frequent trades. Memecoin rug pulls: run a rug check before a buy. Key and skill security: AI agent security. Running money for other people is regulated in most places. This guide is about your own money only.
Contests where models trade live, such as Alpha Arena, are a different setup. Moby Agents do not trade perps. Whether the whole idea is a product or a pitch is covered in is AI trading legit. Building the bot yourself is how to build an AI trading bot. Starter plans are on strategy ideas.
Frequently asked questions
What is an AI hedge fund?
It is an investment setup where AI models help choose and size trades. The term covers professional funds that use machine learning, open-source agent teams that simulate a fund for research, and personal setups where one person lets an AI agent trade a small budget under limits.
Is the ai-hedge-fund GitHub project real trading?
No. Its README calls it a proof of concept for educational purposes and says the system does not actually make any trades. It is useful for learning how investor-style agents analyze tickers, and it includes backtesting and paper trading modes.
What is TradingAgents?
TradingAgents is an open-source multi-agent framework that mirrors a trading firm, with analyst, researcher, trader, risk and portfolio manager agents. Per its README, approved orders go to a simulated exchange, and the project is designed for research purposes.
Can an AI agent manage my portfolio?
An AI agent can research and place trades if you connect it to an account or wallet that allows it. Keep the budget small, use a separate wallet, enforce limits outside the model, set exits on every trade and review its activity. It is not a substitute for professional advice.
How do I start a personal AI fund onchain?
Write down what you will trade, your position size, daily budget and exits. Use research agents for ideas, then give one execution agent its own wallet with a small balance and server-side limits. Log every trade and review it weekly.
Can a Moby Agent run several strategies at once?
No. Each user gets one agent, with its own wallet separate from your main in-app wallet. You can do research in other tools and let your one Moby Agent execute, with your limits checked server-side on every buy.
Do AI hedge funds make money?
There is no general answer. Results depend on the strategy, data, fees and market, and backtests often look better than live trading. Treat any AI fund, including your own, as risky and size it so a loss would not hurt you.
Try it with a Moby Agent
Set up a Moby Agent with its own wallet and your spending guardrails, then connect it to the AI you use.
You set your agent's limits on max per trade, daily buy volume and market-cap range from the Moby app, and Moby checks them server-side on every buy your agent makes.
Moby Agents trade Solana tokens, paid in USDC, including tokenized stocks.
Each user gets one agent, with its own wallet separate from your main in-app wallet. The agent cannot withdraw or send funds.
Connect it from Codex, Claude Code (CLI), Gemini CLI (with a paid Gemini API key or enterprise access) and Grok CLI, with the agent key. In ChatGPT, paste your personal Moby connector URL from the app. ChatGPT: works on Business, Enterprise and Edu workspaces that allow custom MCP servers (web).
Tell your agent your exits and it places real take-profit and stop-loss orders on Moby, including a one-call bracket.
A 0.75% swap fee (minimum $0.90 per swap), plus onchain fees.
10% of the Moby fee back in $MOBY on every swap, 20% after your first swap, up to 30% as you climb tiers.
Start at Moby Agents. The product frame is agentic trading.
Trading crypto and tokenized stocks is risky and you can lose money. Nothing on this page is investment advice. Facts about other products are from their own sites as of November 3, 2026; check them for current details.