VectorBT
vectorbt.dev
Python library for vectorized backtesting that tests thousands of strategy variations at once with NumPy and Numba; open-source core, paid PRO membership.
What it is
VectorBT, created by Oleg Polakow, is a Python package for research-style backtesting. Instead of looping through bars one strategy at a time, it represents many strategy instances, instruments and time periods as arrays and runs them together, compiled with Numba. That makes parameter sweeps, such as every pair of moving-average lengths across a basket of symbols, fast enough to run in a notebook. Results come with portfolio statistics and interactive Plotly charts that work in Jupyter.
The open-source version installs with pip and is free under Apache 2.0 with the Commons Clause, which allows personal and business use but not selling a product that is mainly VectorBT itself. VectorBT PRO is a separate, faster-moving private codebase offered as a membership: it adds parallelization, portfolio optimization, pattern recognition, limit orders, leverage and many other features, plus private documentation and a Discord community. The provider treats the open-source version as the free trial for PRO.
This is a library, not an application. You need to be comfortable with Python and pandas, bring or download your own data, and write your own strategies; there is no point-and-click interface and no live broker execution in the open-source core.
Key features
- Vectorized backtesting
- Runs many strategy configurations, instruments and periods in one pass using NumPy arrays compiled with Numba.
- Parameter sweeps
- Test large grids of indicator settings at once and compare their performance side by side.
- Indicators and signals
- Built-in indicators and signal generators, with support for custom indicators.
- Portfolio statistics
- Returns, drawdowns, trade records and performance metrics for each simulated portfolio.
- Interactive charts
- Plotly figures and Jupyter widgets for exploring results in a notebook.
- Optional Rust backend
- Install the rust extra for additional speed-ups alongside Numba.
Pricing
| Plan | Price | What you get |
|---|---|---|
| VectorBT (open source) | $0 | pip install vectorbt; Apache 2.0 with Commons Clause |
| VectorBT PRO monthly | From $25/mo | Private repository, documentation and Discord; cancel anytime |
| VectorBT PRO 12 months | From $240/yr | One payment for a year of access |
| VectorBT PRO lifetime | From $500 one time | Permanent access including updates |
PRO individual memberships are for personal, non-commercial use; commercial use needs a separate licence arranged with the developer. If a PRO membership ends, the last installed version keeps working but updates, the repository and support stop.
Free option: The open-source library is free to install and use, and the provider treats it as the trial for PRO.
Checked September 25, 2026 on vectorbt.pro. Prices change; confirm on the provider's site before you subscribe.
Strengths
- Tests thousands of parameter combinations far faster than loop-based backtesters
- Free open-source core usable by individuals and businesses
- Works inside Jupyter with interactive Plotly charts
- A PRO membership with a lifetime option for traders who outgrow the free version
Limitations
- Requires Python and pandas skills; there is no graphical interface
- No bundled market data or live trading in the open-source core
- Many features, including portfolio optimization, limit orders and leverage, are PRO-only
- PRO individual memberships exclude commercial use
- Vectorized testing handles path-dependent logic less naturally than event-driven backtesters
Good fit for
- Python quants sweeping large parameter spaces
- Researchers comparing many signals across many symbols
- Traders who already work in Jupyter and pandas
Look elsewhere if
- Traders who want a no-code backtester
- Anyone who needs backtesting and live execution in one package (an event-driven framework or platform fits better)
Compare
Alternatives to VectorBT
VectorBT questions
Is VectorBT free?
Yes. The open-source vectorbt package is free on PyPI under Apache 2.0 with the Commons Clause. VectorBT PRO is a paid membership.
How much does VectorBT PRO cost?
Memberships start at $25 a month, $240 for 12 months, or $500 one time for lifetime access.
What does VectorBT PRO add?
Parallelization, portfolio optimization, pattern recognition, limit orders, leverage and many other features, plus private docs and a Discord community.
Is there a VectorBT PRO free trial?
No separate trial. The developer points to the free open-source version as the way to try the workflow first.
Can VectorBT trade live?
The open-source library is for research and backtesting; it does not place orders with a broker.
Written from VectorBT's own website and documentation. All Trading Tools lists and compares tools; it does not recommend trades or give financial advice. Sources: vectorbt.dev, vectorbt.pro, vectorbt.pro/become-a-member, vectorbt.pro/faq, github.com/polakowo/vectorbt.