The full list of 10 free, open-source AI trading and finance repos — links, setup commands, and what each actually does — so you can clone one and run it today.
This guide is reviewed for clarity, service accuracy, and AI-search readability. The next quarterly content review is tracked internally before unsupported metrics or client proof are added.
The promise
Ten free, open-source repos that give you AI-assisted trading and finance research — no Bloomberg subscription, no hedge fund payroll. Below: what each one does, the GitHub link, the exact commands to get it running, and how hard the setup actually is.
One note before you start: these projects update fast. If a command below doesn't match what you see after cloning, the repo's README is always the source of truth for the current entry-point script — check it first.
Quick-pick table
| # | Repo | What it does | Setup |
|---|---|---|---|
| 1 | AutoHedge | 4-agent automated hedge fund pipeline | Medium |
| 2 | Vibe-Trading | 64-skill agent debate engine for crypto/stocks/options | Medium |
| 3 | Fincept Terminal | Free Bloomberg Terminal alternative | Easy |
| 4 | TradingAgents | Multi-agent trading firm simulation | Medium |
| 5 | AI Hedge Fund | 14 investor-persona agents debate stock picks | Medium |
| 6 | FinRL | Reinforcement learning for trading | Advanced |
| 7 | FinRobot | LLM agents for equity research | Medium |
| 8 | AI-Trader | Autonomous trading agent framework | Medium |
| 9 | FinGPT | Open financial LLM | Advanced |
| 10 | Qlib | Microsoft's quant investment platform | Medium |
1. AutoHedge — a hedge fund in one command
AutoHedge runs an automated fund out of four specialized agents that hand work to each other in sequence:
- Research agent — pulls data and forms a thesis on a ticker
- Validation agent — stress-tests that thesis against the numbers
- Risk agent — sizes the position and sets limits
- Execution agent — places (or simulates) the trade
Repo: github.com/The-Swarm-Corporation/AutoHedge
bashgit clone https://github.com/The-Swarm-Corporation/AutoHedge.git cd AutoHedge pip install -r requirements.txt python main.py
You'll need an LLM API key (Anthropic or OpenAI) set as an environment variable before the four agents can reason — the repo's .env.example shows which keys it expects.
2. Vibe-Trading — agents that argue before they trade
Vibe-Trading ships 64 built-in finance skills — technical analysis, fundamental screening, sentiment scoring, options-pricing math, and portfolio risk checks — as callable tools its agents can invoke.
The core idea: multiple agents each run a different subset of those 64 skills against the same setup, then debate the result in a structured back-and-forth (bull case vs. bear case vs. risk check) before a final decision agent commits to a trade. It works across crypto, stocks, and options, so you're not locked into one asset class.
Repo: github.com/HKUDS/Vibe-Trading
bashgit clone https://github.com/HKUDS/Vibe-Trading.git cd Vibe-Trading pip install -r requirements.txt python main.py
3. Fincept Terminal — Bloomberg's $24,000/year replacement
A Bloomberg Terminal subscription runs about $24,000 a year. Fincept Terminal is a free, open-source desktop app aiming at the same job: 20+ AI agents modeled on named investors (Buffett, Soros, Dalio among them) analyze markets alongside CFA-style analytics, pulling from 100+ data sources — no per-seat license required.
Repo: github.com/Fincept-Corporation/FinceptTerminal
bashgit clone https://github.com/Fincept-Corporation/FinceptTerminal.git cd FinceptTerminal pip install -r requirements.txt python app.py
7 more repos to round out your stack
4. TradingAgents
Multi-agent LLM framework that simulates an entire trading firm — analyst desks, a research team, and a trader agent all passing information up the chain before a decision. 80k+ GitHub stars.
Repo: github.com/TauricResearch/TradingAgents — Setup: Medium
bashgit clone https://github.com/TauricResearch/TradingAgents.git cd TradingAgents pip install -r requirements.txt python main.py
5. AI Hedge Fund
14 legendary-investor personas (Buffett, Munger, Burry, and others) plus analytical agents debate a stock pick until they converge on a call. 59k+ GitHub stars.
Repo: github.com/virattt/ai-hedge-fund — Setup: Medium
bashgit clone https://github.com/virattt/ai-hedge-fund.git cd ai-hedge-fund poetry install poetry run python src/main.py --ticker AAPL
6. FinRL
Deep reinforcement learning framework for training trading agents from historical data — the RL model learns a policy instead of following hand-written rules. 15k+ GitHub stars. Mostly notebook-driven; a GPU speeds up training but isn't required for the demo notebooks.
Repo: github.com/AI4Finance-Foundation/FinRL — Setup: Advanced
bashgit clone https://github.com/AI4Finance-Foundation/FinRL.git cd FinRL pip install -r requirements.txt
Then open and run the example notebooks in the repo's examples/ folder — that's how FinRL demonstrates training and backtesting.
7. FinRobot
LLM agent platform from the same team as FinRL, aimed at automating equity research — pulling filings and market data, then drafting analyst-style reports.
Repo: github.com/AI4Finance-Foundation/FinRobot — Setup: Medium (star count moves fast — check the repo for the live number)
bashgit clone https://github.com/AI4Finance-Foundation/FinRobot.git cd FinRobot pip install -r requirements.txt python main.py
8. AI-Trader
Autonomous trading agent framework from the HKUDS research group (same lab behind Vibe-Trading) — an agent loop for market analysis and trade signal generation.
Repo: github.com/HKUDS/AI-Trader — Setup: Medium
bashgit clone https://github.com/HKUDS/AI-Trader.git cd AI-Trader pip install -r requirements.txt python main.py
9. FinGPT
An open-source financial LLM, fine-tuned for tasks like sentiment analysis on financial news and earnings calls. Runs as a model you load via Hugging Face transformers rather than a standalone app.
Repo: github.com/AI4Finance-Foundation/FinGPT — Setup: Advanced (GPU recommended for inference)
bashgit clone https://github.com/AI4Finance-Foundation/FinGPT.git cd FinGPT pip install -r requirements.txt
Run the demo notebook in the repo to load a fine-tuned checkpoint and test it against sample text.
10. Qlib
Microsoft's quant investment platform — data pipelines, factor research, and model backtesting in one framework. The most "institutional" tool on this list.
Repo: github.com/microsoft/qlib — Setup: Medium
bashpip install pyqlib python -m qlib.run.get_data qlib_data --target_dir ~/.qlib/qlib_data/cn_data --region cn qrun examples/benchmarks/LightGBM/workflow_config_lightgbm_Alpha158.yaml
Before you clone any of these
These are research and automation tools, not financial advice. Backtest results are not future returns, and every agent above can be confidently wrong — treat their output as a second opinion, not an execution order.
If you're building a workflow like this into a product rather than just running it for yourself, that's exactly the kind of AI tooling Hypd tracks and builds with for solo founders — worth a look if you want the next layer beyond "run it locally."
FAQ
Do I need to know how to code to use these?
Yes, at a basic level. All ten are Python repos you run from a terminal — comfort with git clone, pip install, and editing a config file is the bar. None have a polished no-code UI yet, though Fincept Terminal comes closest with its desktop app.
Which one should I start with if I've never touched an AI trading agent before? Fincept Terminal — it's the easiest install and gives you a working dashboard to explore before you touch any agent logic. Once you're comfortable, AutoHedge or AI Hedge Fund are the next step up for seeing multi-agent decision-making in action.
Can I connect these to a real brokerage account? Some can — AI Hedge Fund and FinRL support broker integrations (Alpaca is common) — but every repo here defaults to paper trading or backtesting mode. Get comfortable with simulated results first, and treat live execution as a separate, deliberate decision. </br>