testfol.io and ENSEMBLE: choosing a research workflow

Choose a research tool by the work you need to do and the evidence you can inspect afterward. ENSEMBLE is research software for saved investment allocation models: define your own logic, inspect its implementation, compare runs, compose models, and maintain allocations through connected interfaces.

What testfol.io offers

The testfol.io pricing page, checked October 5, 2026, lists portfolio and tactical backtesting, simulated tickers, ticker modifiers, optimization, and financial-planning tools. It also lists MCP access. Features and limits differ by plan; consult that page and the product's documentation before purchasing.

Earlier versions of this comparison described testfol.io as lacking agent access. That description is no longer current. We have not independently tested every competitor tool or its MCP connection.

Where ENSEMBLE fits

ENSEMBLE supports models defined in natural language or supplied Python that satisfies its runtime contract. Models have declared inputs and stored builds and runs. An isolated parameter test can be compared with another run before you choose whether to adopt it.

A model can hold assets, other models, or both, within composition and runtime limits. Reusable optimizers and explicit backfills let you keep assumptions with the research. Data and dependency changes can affect later runs; stored records are not immutable copies of every upstream input.

Use the web app, a compatible MCP client, REST API, or CLI to work with the same platform. Client permissions and authentication requirements vary. See the portfolio walkthrough for a first model.

Compare equivalent results

Match the asset universe, dates, dividend treatment, signal timing, rebalance rules, and execution assumptions before comparing return figures. Check whether a period uses synthetic history and whether costs are actually modeled. ENSEMBLE does not universally deduct transaction costs from every model; inspect the script and selected result view.

For a leveraged ETF, confirm that the instrument and required history are present in the catalog. A backfill or custom leverage model is an assumption about earlier behavior, not observed performance of a fund before its inception.

Costs and ongoing use

Consult ENSEMBLE pricing for subscription tiers, starting-credit terms, and metered builds and runs. Reads are free. Enabled schedules depend on readiness, data availability, and account limits. Business adds broker-format files and other distribution features; exports do not place trades or establish broker tradability.

Try the same small, fully specified research question in each product. Inspect how the result was produced, how revisions are retained, and what maintaining it would cost. A hypothetical backtest cannot establish which allocation you should hold or what it will return. The products are independent and unaffiliated.