Composer Trade Review: Pricing, Fees and Backtest Limits

Composer is worth paying for if you want to run rules-based strategies automatically, you are comfortable moving money into a new brokerage account to do it, and your strategy is a weighting rule rather than an entry and exit rule on one stock. The Starter plan is $0 and automates your first strategy, so the honest answer for most people is that you should not pay anything yet. The two facts that decide the rest are not in any ranking review: Composer was acquired by SoFi and relaunched as Composer by SoFi on June 23, 2026, and the trade timing that decides whether your live results match your backtest is itself a paid feature.

Everything below was read from Composer's own pricing page, its knowledge center and SoFi's press release, in September 2026. Where the vendor's own pages contradict each other, the review says so.

Composer at a glance

Is Composer worth it?

Start with the plan that costs nothing, because it is more generous than the reviews suggest. Starter automates one strategy, gives you the no-code editor, lets you backtest without limit, and lets you save, watch and share strategies. Composer's own line is "No subscription needed to start automated trading". If you have a single allocation rule you want running on autopilot, you can do that for $0 and the only money involved is the money you deposit.

You start paying when you want more than one strategy running, or when your strategy gets structurally bigger. That second axis is the one nobody explains. Composer prices by nodes, which are the individual blocks in a strategy's decision tree. Starter tops out at 200 nodes, Advanced at 350, Pro at 10,000 and Business at 15,000. A simple "if SPY is above its 200-day, hold SPY, otherwise hold treasuries" rule uses a handful. A nested momentum rotation that checks six conditions across twelve tickers can run into the hundreds fast. So the jump from Starter to Advanced is usually triggered by strategy complexity rather than by ambition.

Composer pricing in 2026

Two things about this table are worth pausing on.

The first is that Pro is annual. The pricing page prints "Billed yearly" above "$32/mo." and "$384/yr" beneath it. There is a question in Composer's own FAQ titled "Can I upgrade from a monthly to a yearly Pro plan?", so a monthly Pro presumably exists somewhere in the account flow, but it is not priced on the public page. Trustpilot reviewers complain about annual membership fees they did not expect, and an annual-only price shown as a monthly figure is how that happens. Confirm the billing term at checkout.

The second is the Business tier's feature list, which includes "Full IP ownership of strategies". That phrase only means something if the individual tiers do not carry it. If you are building strategies you consider proprietary, read the terms of service before you build them on a Starter account. That is not an accusation, it is a line item on the vendor's own pricing page that deserves a direct question to support.

The Pro feature that should worry you on the free plan

Here is the finding that changed how I read Composer's pricing. Under Pro, and only under Pro, is this line: "Get priority automated trade timing closer to market close".

Now read Composer's knowledge center on how backtests work. Backtests use daily adjusted closing prices. And the knowledge center warns that live trading uses real-time data, so "decisions made during this window may differ from those suggested by a backtest due to volatility".

Put those two together. Your backtest assumes you traded at the close. How close to the close you actually trade is sold as an upgrade. On a strategy that rebalances rarely, this is noise. On a daily momentum rotation into leveraged ETFs, which is exactly what a large share of Composer's community strategies are, the gap between a 3:45 pm fill and a 3:58 pm fill is a real and recurring cost that your free-plan backtest never showed you. The cheapest plan runs the strategies most sensitive to the thing the cheapest plan does not give you.

How good is Composer's backtesting?

Better than I expected, and more honest than most. The data goes back to 1990 on daily adjusted closing prices that account for dividends and splits, which is a longer and cleaner record than several paid backtesters offer. Composer models regulatory trading fees and applies slippage, and its documentation states plainly that "all backtest results are hypothetical and benefit from hindsight" and that different platforms use different assumptions, so cross-platform comparisons are unreliable. That last point is unusually candid for a vendor.

Two settings deserve your attention before you trust a curve.

The default slippage is 1 basis point. That is one hundredth of one percent. On a $10,000 rebalance it is a dollar. For SPY at midday that is roughly fair. For a small-cap ETF, a leveraged product, or any trade you are placing into the closing minutes, it is optimistic by a wide margin. The setting is adjustable, so adjust it. Re-run your strategy at 5 and 10 basis points and see whether the edge survives. If it does not, you found that out for free.

The subscription cost built into the model is stale. Composer's backtest documentation offers to include "a Composer subscription ($40/month)" in the modeled costs. There is no $40 plan on the current pricing page. The plans are $0, $10 and $32 billed yearly. That is a small inconsistency between the help center and the pricing page, and it is the kind of thing to notice because it tells you which page was updated when the plans changed and which was not.

The deeper limitation is structural rather than a setting. A backtest on Composer is built in the same editor that builds the live strategy, which is convenient and also means the test inherits every assumption the builder made. Nothing in the tool argues back. We wrote up the failure modes this produces in common backtesting mistakes, and the overfitting one is the killer here: with 2,000 community strategies visible and a free unlimited backtester, the temptation is to browse curves until one looks good, which is search, not research.

What the SoFi acquisition changes

SoFi completed its acquisition of Composer Securities LLC in April 2026 and relaunched the product as Composer by SoFi on June 23, 2026. SoFi's press release describes an "AI-powered investing platform that helps investors move from investing ideas to automated execution in minutes", cites over 2,000 community-built strategies, and quotes CEO Anthony Noto saying Composer "has built one of the most innovative AI-powered investing platforms available to retail investors today". SoFi says Composer's technology will be integrated into the wider SoFi platform and expanded to SoFi members through SoFi Plus.

For a buyer, three practical consequences follow.

The counterparty risk profile improved. You are now depositing with a subsidiary of a listed company with 14.7 million members rather than with a venture-funded startup. For anyone who hesitated to move real money into an independent algo platform, that is the objection answered.

The pricing is new and will probably move again. The site currently runs a banner reading "Get Composer for $0/month. New pricing plans + new referral program." Pricing that recently reset under new ownership, with a referral program attached, is pricing that is still being calibrated. Every review published before mid-2026 describes a different ladder.

And the marketing got louder while the caveats got quieter. The June press release says strategies "can be tested against historical data" and that investors can "review how it would have performed historically". It does not address the limitations of backtested results; its cautionary statement covers forward-looking business statements about product adoption and market conditions. Composer's own knowledge center is considerably more careful about hindsight than its parent's launch announcement is. Read the knowledge center.

Who Composer is right for, and who it is not

The entity case deserves a line of its own, because the Business tier is the only one that mentions mark-to-market accounting and it is easy to read that as a nice-to-have. It is not. If you are trading through an LLC or a partnership, the algorithms generate the transaction record and somebody still has to turn that record into a P&L, a balance sheet and a cash-flow statement that a partner, a lender or an accountant will accept. Plenty of small trading entities do that from a bookkeeping export and generate the statements separately, which is usually cheaper than buying a tier for the accounting feature alone. Price the two routes before you book the sales call.

Composer alternatives worth pricing against it

The honest comparison set is small, because very few products combine a no-code strategy builder with actual execution.

Portfolio Visualizer does the allocation math without touching your money, which suits people who want the study and not the automation. Its free tier caps portfolios at 15 assets with limited history, and the paid plans are $30 and $55 a month billed annually. We broke the whole ladder down on our Portfolio Visualizer cost page, including the licence line that decides which plan a professional is allowed to buy.

QuantConnect is the answer if you can write code and want institutional-grade data and event-driven backtesting. It is a genuinely different league of rigor and a genuinely steeper wall. Our QuantConnect alternative page covers where each side of that trade makes sense.

TradingView handles the signal and charting half well and the execution half through third-party bridges, which introduces a failure mode Composer does not have. A webhook between an alert and a broker can stop working silently, and nobody tells you, so if you go that route the bridge needs its own monitoring rather than an assumption that no alerts means no signals. Our TradingView backtesting page covers what its strategy tester can and cannot do.

And the thing none of them do, including Composer, is argue with you. That is the gap we built our own tool for: you describe an idea in plain English, it gets restated as a rule you can actually test, it runs on adjusted daily history, and you get a verdict that says the idea did not work when the idea did not work. It does not execute trades and it does not hold your money, so it is not a Composer replacement for the automation half. It is the second opinion before you automate anything.

The verdict

Composer is a good product that is now owned by a serious company, and its free tier is generous enough that paying before you have outgrown it would be a mistake. Advanced at $10 a month is fair value for anyone running a handful of rules. Pro at $384 a year is worth it for two reasons that have nothing to do with the 500-strategy headline: API access, and trade timing closer to the close.

The thing to be careful about is the one the platform makes easiest. A free unlimited backtester sitting next to 2,000 community strategies is an invitation to shop for a curve. Raise the slippage above the 1 basis point default, test your rule on a window you did not pick after seeing the result, and find out whether the idea survives contact with costs before you fund the account.