Portfolio123

What it is

Portfolio123 is a hosted research platform for rules-based stock selection. The working unit is a ranking system: a weighted tree of factors — value, growth, quality, momentum, estimate revisions — that scores every stock in a universe on each rebalance date. A screen filters on it, a simulation buys the top of the list and sells by your rules, and a "book" combines several simulations. All of it is built in forms and a formula language, with no Python required; the vendor counts 4,300+ factors and 430+ functions.

That is why it sits with the backtesting tools. People pay for the point-in-time history underneath — would buying the 20 best-ranked stocks every week have worked over fifteen years? — and the plan names say so: Screener, Backtest, Portfolio, Ultimate, priced mainly by how many years you can test over. The screener sells on its own as the entry plan, hence the second listing among screeners.

AI Factor, on the Portfolio and Ultimate plans, trains machine-learning models on factors you choose and turns their predictions into a ranking you can backtest. Training and predictions carry usage fees on top.

Pricing

A free account and four paid plans, in US dollars, with a separate and higher price list for professional users. The monthly figures above are month-to-month; annual billing is three months free on the retail plans (save up to 28%) and two on the professional ones (up to 23%).

A plan covers one region — North America (US and Canada) or Europe and the UK — and a second region adds 40%. Limits scale with the tier: resource units run 400, 1,000, 5,000 and 10,000; API credits 250, 500, 1,000 and 10,000 a month, which do not roll over, with more sold on demand. Strategy positions, linked broker accounts and optimizer combinations are capped too.

Data & coverage

Fundamentals and estimates come from FactSet and prices and volume from ICE, per the vendor's data article. Delisted stocks stay in the history, so backtests are not survivorship-biased. Stock history goes back 5, 10, 15 or 20 years depending on the plan; ETF price history runs from 1999 on every plan. The research engine runs on end-of-day prices. The Manage section shows free IEX real-time prices, falling back to delayed SIP prices and then to the previous close for names IEX does not quote.

Exchanges covered include NYSE, Nasdaq, NYSE Arca, NYSE American, OTC, TSX, TSX Venture, the London Stock Exchange, Xetra, the Euronext markets, SIX and the Nordic exchanges; the Europe region's country list runs to 39 countries, Turkey and Ukraine included. Stocks and ETFs only: no options, futures, forex or bonds.

Integrations

The DataMiner API runs screens, ranks, strategies and AI Factor jobs, and uploads your own factors, metered in API credits per call. Portfolio123 publishes the p123api Python wrapper on PyPI (3.1.0, uploaded 25 August 2026). DataMiner is the no-code front end to the same calls: batches of operations written as human-readable instructions.

Broker links go to Interactive Brokers and Tradier. Holdings and fills sync back into Manage, and a rebalance can be sent as orders in one step.

Limitations

• Your results do not leave the platform freely. Raw factor values are licensed. Without your own FactSet or Compustat licence, screen downloads are truncated to 500 tickers for recent dates and blank for anything ten days old or more — so the history you test on is not data you can take away.
• Point-in-time is approximate. The vendor says FactSet backfills figures after an announcement, so re-running a backtest later can change its result.
• Stocks and ETFs on daily data. No intraday testing, no derivatives, no coverage of Asia, Australia or Latin America.
• Two brokers. Automated execution means Interactive Brokers or Tradier; anyone else is manual.
• A model marketplace sits beside the tools. Portfolio123's own Models and third-party Designer Models deliver ongoing trade recommendations from someone else's rules; Designer Models are paid for directly, with the designer keeping 65%. That is a product for following another person's selections, separate from the research software this card is about.

Alternatives

QuantConnect is the code-first route to the same question, with more asset classes and full control of the engine. Stock Rover and Finviz Elite are cheaper if screening is all you need and a backtest is not. Composer is no-code too, but tests ETF and stock allocation rules on price data rather than ranking stocks on fundamentals. AmiBroker with Norgate Data is the desktop pairing for price-based systems on survivorship-free history.