QuantConnect or your own trading stack?
QuantConnect is one of the most capable research and backtesting platforms available, and its engine, LEAN, is open source. That makes the question less “QuantConnect or not” and more “how much of the stack do you want to run yourself”. Here are the three options.
The short answer
For a research-heavy strategy on a supported broker, QuantConnect’s cloud is a strong option and often the fastest. Run its open-source engine, LEAN, on your own servers if you need control over where the system runs or what data it uses. Choose a custom system when you need a broker LEAN does not support, several brokers or accounts under one risk limit, or close integration with your own databases.
Option 1: QuantConnect’s cloud
You write the strategy in Python or C#, research and backtest it on QuantConnect’s cloud, and deploy it live to a supported brokerage. LEAN models fees, slippage, spreads and margin, handles portfolios across many asset classes at once, from equities and options to futures, forex and crypto, and moves from research to backtest to live trading with little change to the code. For a research-heavy strategy on a supported broker, this is a strong option and often the fastest one.
Option 2: LEAN on your own infrastructure
Because LEAN is open source, you can run the same engine on your own servers, with your own data and brokerage connections, and avoid being tied to one provider. This suits traders and firms who like LEAN’s design but need control over where the system runs or what data it uses. Where LEAN lacks a connection you need, it can be extended with a custom brokerage or data integration.
Option 3: a system built for you
A custom system on your broker’s API makes sense when the strategy needs something a general engine does not provide comfortably: a broker LEAN does not support, several brokers or accounts under one risk limit, unusual order handling, or close integration with your own databases and dashboards. You own every line, and the system does only what your strategy needs. See custom trading software and ourbroker API comparison.
Side by side
| QuantConnect cloud | Self-hosted LEAN | Custom system | |
|---|---|---|---|
| Languages | Python or C# | Python or C# | Whatever fits the broker and latency budget |
| Brokers | QuantConnect’s supported list | Supported list, or a custom integration | Any broker with an API |
| Research and backtesting | Built in, on cloud hardware | Built in, on your hardware | Built to the strategy, or LEAN used for research |
| Where it runs | QuantConnect’s cloud | Your servers | Your servers or cloud |
| Ownership | Your strategy code, their platform | Your code and your deployment of open-source LEAN | All of it |
A common combination
Many traders research on QuantConnect and then decide where the live system should run. That works well, as long as the live system is checked against the research version on the same data before it trades, so differences in fills, fees and data do not go unnoticed. Our guide to the edge cases of live trading systems covers what changes between a backtest and a live market. If your live account is at Interactive Brokers, see our Interactive Brokers integration page.
Common questions
- Can I run QuantConnect’s engine on my own server?
- Yes. QuantConnect’s engine, LEAN, is open source, and QuantConnect itself notes that you can run it with your own infrastructure, data and brokerage connections instead of on their cloud.
- Does QuantConnect support my broker?
- LEAN integrates with a range of brokerages and data providers; check QuantConnect’s current list for yours. If your broker is not supported, the options are a custom brokerage connection for LEAN or a system built directly on your broker’s API.
- Which languages does QuantConnect use?
- Strategies are written in Python or C#. The engine itself is written in C# and runs on Linux, Mac and Windows.