LumiBot indicators compared¶

How do custom indicators in LumiBot compare with the tools traders already use? The tables below cover what each one lets an indicator author do. Facts were checked on October 25, 2026 from each product’s own documentation (sources below). Products change; check their docs before you decide.

Language and where it runs¶

Product	Language	Where it runs	Open source
LumiBot	Python 3	Your computer or server	Yes, GPL-3.0
TradingView (Pine Script)	Pine Script v6	TradingView’s servers	No
LuxAlgo Vela + PineTS	Pine v5/v6 run by PineTS (TypeScript)	The viewer’s browser (or a server)	Vela core Apache-2.0; PineTS AGPL-3.0
OpenMarket (wrun)	One TypeScript file compiled to WebAssembly	Browser; protected code on OpenMarket’s servers	No
NinjaTrader 8	NinjaScript (C#)	NinjaTrader desktop app	No
MetaTrader 5	MQL5	MetaTrader terminal	No
QuantConnect LEAN	Python or C#	Locally or QuantConnect’s cloud	LEAN engine, Apache-2.0

What an indicator can do¶

Product	Any library (numpy, pandas, ML)	Call an outside web API	Same code places live broker orders
LumiBot	Yes	Yes (Using outside data in an indicator)	Yes, inside a strategy on your broker
TradingView (Pine Script)	No, Pine built-ins and Pine libraries	No	Through alerts and webhooks to other services
LuxAlgo Vela + PineTS	No, PineTS built-ins	No; the chart feeds the data	No (paper trading)
OpenMarket (wrun)	No imports	Built-in feeds only	No (“no orders from the file”)
NinjaTrader 8	No Python. C# on .NET Framework 4.8; third-party .NET DLLs can be referenced	Technically possible with standard .NET classes; NinjaTrader support calls it undocumented and unsupported	Yes, NinjaScript strategies
MetaTrader 5	No Python libraries. MQL5 (C++-like); DLL imports if the user turns them on	No. `WebRequest()` is blocked in indicators (error 4014). It works only in Expert Advisors and scripts, for URLs the user allow-lists, and never in the Strategy Tester	Yes, Expert Advisors
QuantConnect LEAN	Yes, several machine-learning libraries in Python	Yes, in the algorithm: `self.download` (with headers or a login) and custom data streams. In their cloud each download times out after 10 seconds and the number of remote files per backtest depends on the plan (Free: 25)	Yes, LEAN algorithms

Live updates and backtests¶

Product	Live (forming) bar	One-bar updates, no full recompute	Backtest the same indicator
LumiBot	You choose: closed bars, or every update with rollback	Yes, built in	Yes
TradingView (Pine Script)	Indicators recalculate on every live update, with rollback	Yes (Pine’s execution model)	Strategy Tester
LuxAlgo Vela + PineTS	Every streamed update	Yes (Pine’s execution model)	Backtests through the hosted app’s AI assistant
OpenMarket (wrun)	Every update	Not covered here	Chart replay
NinjaTrader 8	`Calculate`: on bar close (default), each tick, or price change	Yes (`OnBarUpdate`)	Strategy Analyzer
MetaTrader 5	`OnCalculate` runs when price data changes	Yes (`prev_calculated`)	Strategy Tester
QuantConnect LEAN	`update()` per data point	Yes (`update`)	LEAN backtests

What this means in plain words¶

If you want Python, real machine learning and outside data, and the same code to trade your own broker, LumiBot and QuantConnect LEAN are the two Python options. LEAN is the closest match: also Python, also open source, also backtests and live trading. In LEAN, outside data is loaded by the algorithm and its cloud has download limits; in LumiBot an indicator calls the API itself, updates the forming live bar with rollback, can stream live in a plain Python script, and runs on your own machine with pip install lumibot.
NinjaTrader indicators are C#. Web calls are possible with .NET classes but NinjaTrader does not support them. MetaTrader 5 blocks web requests in indicators entirely.
If you want the biggest charting community, TradingView is it, with Pine Script’s limits (no outside APIs, no Python libraries).
Vela gives you a fast open-source chart in the browser that runs Pine through PineTS. It does not run Python and scripts cannot fetch outside data.
NinjaTrader and MetaTrader are desktop platforms with their own languages (C# and MQL5) and their own brokers and data connections.

Nothing here says any tool makes money. Indicators describe the past; test any idea on paper first.

Sources¶

Checked October 25, 2026.

TradingView Pine Script v6 manual, “Execution model”: https://www.tradingview.com/pine-script-docs/language/execution-model/
LuxAlgo Vela (license and architecture): https://github.com/LuxAlgo/vela ; PineTS (license): https://github.com/LuxAlgo/PineTS
OpenMarket wrun documentation: https://openmarket.xyz
NinjaTrader Calculate property: https://docs.ninjatrader.com/ninjascript/calculate
NinjaTrader outside calls and third-party DLLs (NinjaTrader support forum and help guide): https://forum.ninjatrader.com/forum/ninjatrader-8/indicator-development/1134931-dynamically-adding-property-values-to-a-script-via-file-api ; https://forum.ninjatrader.com/forum/ninjatrader-8/add-on-development/1123485-how-do-i-make-an-external-call ; https://ninjatrader-devel.ninjatrader.com/support/helpGuides/nt8/using_3rd_party_indicators.htm ; https://forum.ninjatrader.com/forum/ninjatrader-8/indicator-development/1207367-trying-to-add-htmlagilitypack-run-into-number-of-errors
MQL5 OnCalculate: https://www.mql5.com/en/docs/event_handlers/oncalculate ; program events: https://www.mql5.com/en/docs/runtime/running
MQL5 WebRequest (not allowed in indicators, not in the Strategy Tester): https://www.mql5.com/en/docs/network/webrequest ; MetaTrader 5 robots and indicators: https://www.metatrader5.com/en/terminal/help/algotrading/trade_robots_indicators
QuantConnect custom indicators: https://www.quantconnect.com/docs/v2/writing-algorithms/indicators/custom-indicators ; LEAN source and license: https://github.com/QuantConnect/Lean
QuantConnect outside data and machine learning: https://www.quantconnect.com/docs/v2/writing-algorithms/importing-data/bulk-downloads ; https://www.quantconnect.com/docs/v2/writing-algorithms/importing-data/key-concepts ; https://www.quantconnect.com/docs/v2/writing-algorithms/machine-learning/key-concepts

Frequently asked questions¶

Questions about what an indicator can do (machine learning, outside data, live values, repainting) are answered on Custom indicators in Python.

Is there a free, open-source alternative to TradingView indicators?¶

Yes. LumiBot is free and open source (GPL-3.0). You write indicators in Python, run them on your own computer, chart them, backtest them and use them in live strategies on your own broker. It is not a charting website; it is a Python library. Install it with pip install lumibot.

Can I write TradingView-style indicators in Python?¶

Yes. A LumiBot indicator declares its settings, outputs, guide lines and whether it draws on the price chart, like a Pine indicator, and processes one bar at a time with Pine-style rollback of the live bar. See Build your first custom indicator.

Do LumiBot’s RSI and EMA match TradingView?¶

The built-ins follow Pine Script v6 formulas, including how averages are seeded and Wilder smoothing for RSI and ATR. Small differences usually come from warm-up history or a different data feed. See Matching TradingView’s numbers.

Can I convert my Pine Script to Python?¶

Yes, by hand or with an AI coding assistant. Matching TradingView’s numbers maps Pine functions to LumiBot ones. Only convert code you wrote or have the right to use; many published scripts have licenses.

Does LumiBot draw charts like TradingView?¶

LumiBot draws indicator charts for backtests. Each indicator’s describe() gives a chart everything it needs (settings, lines, colors, guide lines, price pane or own pane), so you can also plot the values with any charting library.

How is LumiBot different from QuantConnect?¶

Both are open source and both run Python strategies in backtests and live. In LumiBot, an indicator itself can call a web API and give a value for the still-forming live bar, and it runs on your own computer with pip install lumibot. In QuantConnect, the algorithm loads outside data, and their cloud limits download time and the number of remote files per backtest. See the tables above for the sources.