Vibe-Trading is a free, open-source research environment for traders and quantitative hobbyists who want to describe a strategy in plain language and test it against historical data. It is built in Python, released under the MIT license, and includes a Windows build or installer. Its main job is research and backtesting rather than live order execution, which makes it a good fit for people who want to validate an idea before risking money on it.
The project is organized around a set of cooperating AI agents. You state a question or a strategy idea in natural language, and the agents break the task into steps, pull market data, run tests and summarize the results. Because the code is public on GitHub, you can inspect how each step works, which matters for anyone who takes their analysis seriously.
What Is Vibe-Trading and How Does It Work?
Vibe-Trading sits between a chat interface and a quantitative toolkit. A user writes a request, such as asking how a moving-average crossover performed on a given set of assets over several years. The agents interpret the request, decide which data and engines are needed, run the calculations and return a structured answer.
The design separates reasoning from computation. The agents plan and explain, while the backtest engines do the numerical work. That separation makes results easier to audit, because each number can be traced back to a specific engine run and a specific data source.
Who Is Vibe-Trading For?
The most natural audience is a trader or analyst who has ideas but limited time to code each test by hand. Someone who already writes Python can go further, because the code is open and the tool is extensible. Students of quantitative finance may also find it useful as a worked example of how a research pipeline can be structured.
It is less suited to someone looking for a plug-and-play trading machine. The emphasis is on research and backtesting. Anyone expecting a button that places trades automatically will need a different kind of program.
Practical Use Cases for Strategy Research
A common workflow is to test whether a simple idea holds up before committing capital. For example, you might ask the tool to compare a trend-following rule against a mean-reversion rule across several years of data. The output lets you see drawdowns, returns and how the two approaches differ over different periods.
Another use is cross-market comparison. Because the CompositeEngine can combine asset classes under one capital pool, you can examine whether a strategy behaves differently when equities, futures or crypto are traded together. Scheduled runs can then keep an eye on the same question as new data arrives.
Why Backtesting Results Need Careful Reading
A backtest shows how a rule would have behaved on past data. It does not show how the rule will behave next month. Overfitting, where a strategy is tuned until it matches history closely, is one of the most common reasons backtests look better than live performance.
Use the tool to challenge your ideas rather than confirm them. Test on periods you did not use to design the rule, consider transaction costs and slippage, and be skeptical of results that look too smooth. Vibe-Trading gives you the mechanics; the judgment about what the numbers mean remains yours.
Key features
Natural-language research agent
Swarm presets
Extensible skill and tool set
Multiple backtest engines
Fallback market data layer
Scheduled reports
Reusable research artifacts
Supported coins
Vibe-Trading covers several asset classes, including A-shares, US equities, crypto, futures, forex and options, through its set of backtest engines. The cross-market CompositeEngine lets these be combined in a single research run. Check the project documentation for the current list of exchanges and data providers, since coverage depends on the configured data sources.
How to install Vibe-Trading on Windows
- 1
Download from the official source
Get the Windows build or installer from the project's GitHub Releases page at https://github.com/HKUDS/Vibe-Trading, or from the official website if one is linked from the repository. Download only from the official website or GitHub, not from third-party mirrors.
- 2
Verify the file
Compare the SHA-256 checksum of the downloaded file with the value published alongside the release, or check the signature if one is provided. A mismatch means the file should not be run.
- 3
Run the installer
Launch the installer and follow the on-screen prompts. Choose an installation folder you can easily find later, since research artifacts and configuration files will be stored under the project directory.
- 4
Configure data and destinations
Set up your data source preferences and, if you plan to use scheduled reports, configure the email destination. Keep API keys for data providers in the configuration file or environment settings rather than in shared documents.
-
Start the first research session
Open the application, pick a swarm preset or write a question in plain language, and run a small test first to confirm that data loads and results appear as expected.
System requirements
- Windows with an installer or prebuilt build available from the official release page
- Enough free disk space for the application and saved research artifacts
- A working internet connection for market data and optional email reports
How to use Vibe-Trading
Vibe-Trading works best when you start with a narrow question and expand from there. Begin with a single asset and a simple rule so you can check that the output makes sense before testing something more complicated. A small, understandable test is more useful than a large run you cannot interpret.
- Open the application and choose a swarm preset that matches your goal, or start from a blank research prompt.
- Describe the strategy in plain language, including the asset, the time period and the entry and exit logic.
- Review the data sources the agents selected and confirm they match the market you intended to test.
- Run the backtest and examine the summary, including returns, drawdowns and the number of trades.
- Save the run as a versioned artifact so you can compare it with later revisions.
- Test the same rule on a different period that was not used to design it.
- If you want recurring checks, set up a scheduled run and choose an email destination for HTML or PDF reports.
Pros & cons
- Free and released under the MIT license
- Open source, so the logic can be inspected and modified
- Natural-language interface lowers the barrier to running backtests
- Seven backtest engines cover many asset classes
- Automatic fallback between data sources improves reliability
- Scheduled reports and saved artifacts support ongoing research
- Focused on research and backtesting, not automated live trading
- Requires patience to learn how the agents and engines interact
- Results depend on the quality and coverage of the configured data sources
- Development is active, so features and interfaces may change between versions
Is Vibe-Trading safe?
Download the program only from the official GitHub repository or the project's official website. Verify the SHA-256 checksum or signature before running the installer, and never enter a seed phrase, private key or exchange password into any third-party website or chat window. If you connect data provider keys, use read-only permissions where possible. Antivirus software may sometimes flag Python-based tools, so investigate any warning by checking the file hash rather than disabling your protection. Because this is a research tool, it does not hold your funds, but treat any connection to a brokerage or exchange account with the same care you would apply to any financial software.
- Official source only. Download from github.com/HKUDS/Vibe-Trading. Ignore ads and lookalike domains.
- Check the SHA-256. Compare the installer hash with the value the developer publishes before running it.
- Never share your seed phrase. No developer, support agent or website — including CryptApps — will ever ask for it.
Alternatives
Compared with TestMyLogic, which focuses on checking strategies for overfitting, Vibe-Trading places more emphasis on natural-language setup and multi-engine coverage across asset classes. Elinesoft XTester is a simulator and backtester that is more traditional in its workflow. Jesse is a Python framework for strategies and backtests, and it suits developers who want to write strategies directly in code, whereas Vibe-Trading lets you describe them in plain language. Bot platforms such as Gunbot, Moon Bot, ProfitTrailer and HaasOnline TradeServer are oriented toward running live trading, which is a different purpose. OctoBot offers a web interface and trading bot functionality. Choose based on whether your priority is research, live automation or coding control.
Specifications
- Price
- Free
- Type
- Open source
- License
- MIT
- Latest version
- v0.1.18
- Last update
- 30 Oct 2026
- Windows support
- Yes (Windows build or installer)
- Source code
- github.com/HKUDS/Vibe-Trading
Frequently asked questions
Yes. Vibe-Trading is listed as free and released under the MIT license, which permits use, modification and redistribution under its terms. Check the license file in the repository for the exact conditions before reusing the code.
Yes, the project provides a Windows build or installer. Download it from the official GitHub Releases page and verify the file before running it.
Vibe-Trading is oriented toward research and backtesting. It is not primarily a live execution bot, so if you need automated order placement you should look at a dedicated trading bot and review its risks separately.
It is open source, which lets you inspect the code, but safety still depends on where you download it. Use only the official repository or website, check the SHA-256 checksum or signature, and never share seed phrases or private keys with the program or any third party.
A backtest applies a trading rule to historical market data to estimate how it would have performed. It helps you spot weak ideas early, but past results do not guarantee future performance, so treat backtests as a filter rather than a forecast.
Yes, the cross-market CompositeEngine can run strategies that span several asset classes against a shared capital pool. This helps you see how a rule behaves when markets are combined rather than tested in isolation.
A traditional trading bot focuses on executing orders according to a fixed rule, while Vibe-Trading focuses on researching and testing rules before they are used. The two can complement each other, with research happening first and execution handled elsewhere.
Basic use relies on natural-language prompts, so you do not need to write code to start. Knowing Python becomes helpful when you want to add tools, adjust engines or read the source to understand how results are produced.
Verdict
Vibe-Trading is a good choice for traders and researchers who want to test strategy ideas quickly, describe them in plain language and check results across several asset classes. It is free, open source and has a Windows build, which lowers the cost of experimenting. It is less suitable for anyone who wants a turnkey live trading machine. Keep backtests in perspective, verify the download, and check the latest version on the official website before you start.