Repolode
Submit
olaxbt

olaxbt/ai-market-maker

Open-source multi-agent crypto trading system where specialist AI desks generate signals, a Risk Guard can veto trades, and backtests run against buy-and-hold.

View on GitHub
Stars
2,106
Forks
302
Open issues
4
Last commit
3 months ago

Repolode describes open-source projects and does not give investment advice. Trading and investing can lose money: assess the risk yourself and follow the laws where you live before using this project.

What is it?

AI Market Maker is a trading stack for crypto that is structured like a small hedge fund. Specialist AI agents, each acting as a trading desk for technical analysis, macro, news, patterns and on-chain or order-flow signals, feed a signal arbitrator built on LangGraph. Its output goes through a portfolio step and then a Risk Guard, which has veto power before any order is executed.

The project covers backtesting and paper trading. Backtests include performance metrics such as Sharpe and Sortino, a comparison against buy-and-hold, and a full trade ledger. The repository ships pinned historical data with a hash manifest so reruns on the same CSV files are reproducible. Paper trading targets Binance Testnet, and a Hyperliquid adapter operates in dry-run mode. A Futu OpenD adapter covers HK and US stock data with simulated orders.

The stack includes a FastAPI backend, a Next.js dashboard for agent traces and backtest reports, and OpenClaw skill packaging. Running agentic mode requires an OpenAI-compatible or Atlas Cloud API key.

Key features

  • 01Multiple specialist AI agents produce signals that a weighted arbitrator combines into BUY, SELL or HOLD decisions.
  • 02A Risk Guard can veto any proposed trade before execution.
  • 03Backtests report Sharpe, Sortino, profit factor and benchmark comparison against buy-and-hold.
  • 04Pinned historical OHLCV, funding, macro and news data supports offline reruns.
  • 05Paper trading on Binance Testnet, plus a Hyperliquid dry-run adapter and a Futu OpenD stock adapter.
  • 06Next.js dashboard shows agent traces, backtest results and topology.

Use cases

  • Researching multi-agent crypto strategies with reproducible backtests on pinned data.
  • Paper trading an agent-driven strategy on Binance Testnet before risking capital.
  • Studying how a risk veto layer can sit between LLM-driven signals and order execution.
  • Packaging a trading agent as an OpenClaw skill.

Quick start

  1. 1

    Clone the repository

    git clone https://github.com/olaxbt/ai-market-maker.git
  2. 2

    Install uv and Python dependencies (install TA-Lib first, see README)

    pip install uv && uv sync --extra dev
  3. 3

    Copy the env template and set OPENAI_API_KEY or ATLASCLOUD_API_KEY in .env

    cp .env.example .env
  4. 4

    Start the full stack with Docker (requires Docker Desktop), then open http://localhost:3000

    docker compose up --build -d

Things to know

  • Agentic mode requires an LLM API key (OPENAI_API_KEY or ATLASCLOUD_API_KEY); there is no silent fallback.
  • Full setup needs TA-Lib's C library plus Docker Desktop for the platform stack.
  • Licensed under AGPL-3.0, a copyleft license that requires sharing source for networked modifications.
  • Live execution is limited to Binance Testnet paper trading; the Hyperliquid adapter is dry-run only.

Topics

#agentic-trading#crypto#multiagent-systems#backtesting#langgraph#risk-management#paper-trading#claw#openclaw#stock

FAQ

What is ai-market-maker?
Open-source multi-agent crypto trading system where specialist AI desks generate signals, a Risk Guard can veto trades, and backtests run against buy-and-hold. AI Market Maker is a trading stack for crypto that is structured like a small hedge fund. Specialist AI agents, each acting as a trading desk for technical analysis, macro, news, patterns and on-chain or order-flow signals, feed a signal arbitrator built on LangGraph. Its output goes through a portfolio step and then a Risk Guard, which has veto power before any order is executed.
Is ai-market-maker free to use? What license does it have?
ai-market-maker is open source under the AGPL-3.0 license. Check the license terms before using it in commercial products.
Is ai-market-maker still maintained?
Updated within 6 months. The last commit to ai-market-maker was on Aug 20, 2026.
What can I use ai-market-maker for?
Researching multi-agent crypto strategies with reproducible backtests on pinned data. Paper trading an agent-driven strategy on Binance Testnet before risking capital. Studying how a risk veto layer can sit between LLM-driven signals and order execution. Packaging a trading agent as an OpenClaw skill.
How do I install ai-market-maker?
Clone the repository: git clone https://github.com/olaxbt/ai-market-maker.git Install uv and Python dependencies (install TA-Lib first, see README): pip install uv && uv sync --extra dev Copy the env template and set OPENAI_API_KEY or ATLASCLOUD_API_KEY in .env: cp .env.example .env Start the full stack with Docker (requires Docker Desktop), then open http://localhost:3000: docker compose up --build -d
Where can I get ai-market-maker?
The source code and installation instructions are on GitHub: https://github.com/olaxbt/ai-market-maker

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects