ai↗Backtest LabTry demo
Product · as it runs today

The Research Lab,
shown as it works.

These screenshots and the video come from the private Research Lab running its real backtesting engine. To keep the public pages free of licensed exchange data, the market behind them is synthetic and labelled so: Real Backtesting Engine · Synthetic Market Data.

Explore Interactive Demo Watch Product Demo
AI Backtest Lab Research Lab: the run form for a Donchian breakout strategy on a synthetic 4h market, and the automated quality check of the finished run showing return after costs, max drawdown, profit factor and number of trades.
A finished quick-check run. The quality check reads return after costs, drawdown, profit factor and sample size in plain words.
01 / Working capabilities

Real screens, one research loop.

Research Lab run form: strategy file, pair, interval and range preset, strategy parameters, and the fee, slippage, starting capital, risk and leverage policy applied to every run.
Configure a strategy: file, pair, interval, range and parameters. Fees, slippage, capital and risk are fixed by one explicit policy.
Research Lab results: performance after fees, slippage and funding with a buy-and-hold reference, and a candlestick chart with the run's entries and exits.
Results after fees, slippage and funding, with buy and hold as a reference, and the candles and trades exactly as the run saw them. Chart by TradingView Lightweight Charts™.
Research Lab run list with two completed Donchian breakout runs on the synthetic 4h market, and the run details: variants tried for the strategy, the recent 12-month check counter, and Markdown and JSON report downloads.
Every run is kept and counted. Each finished run can be reopened, checked on the recent 12 months or exported as a report.
Research Lab comparison of two parameter variants on the same synthetic data: return, drawdown, profit factor, trade count and both equity curves.
Compare variants on the same data: return, drawdown, profit factor, trades and equity curves side by side.

Charts in the Research Lab and the video are built with TradingView Lightweight Charts™.

02 / Working Local Product Demonstration

One uncut workflow.

One uncut recording of the private Research Lab: configure a strategy, run two real backtests on synthetic market data, read the quality check, performance, chart and trades, then compare the variants.

Working local product demonstration · 76 s, no sound, English captions · Real Backtesting Engine · Synthetic Market Data
Transcript
  1. 0:00The Research Lab runs the real backtesting engine, here on synthetic market data.
  2. 0:04Choose a strategy, a pair, an interval and a range.
  3. 0:08Set parameters. Fees, slippage, capital and risk follow one explicit policy.
  4. 0:14Run a historical backtest. Each run is isolated in its own container.
  5. 0:18The quality check reads return after costs, drawdown, profit factor and sample size.
  6. 0:25Performance after fees, slippage and funding, with buy and hold as a reference.
  7. 0:31Candles and trades exactly as the run saw them.
  8. 0:38The equity curve and every closed trade.
  9. 0:47Test a second parameter variant on the same data.
  10. 0:53Compare variants: return, drawdown, profit factor and equity, side by side.
  11. 1:04The newest 12 months stay held out for a separate recent-period check.
  12. 1:13AI Backtest Lab · aibacktestlab.com
03 / Architecture

From data to a decision you can check.

  1. 01

    Historical market data

    Closed candles, funding and instrument rules, read and pinned for each run.

  2. 02

    Backtesting engine

    Simulates the strategy with explicit fees, funding and slippage, one isolated runner per run.

  3. 03

    Quantitative metrics

    Return after costs, drawdown, profit factor, trades, equity and a buy-and-hold reference.

  4. 04

    Research validation

    Automated quality check, run comparison and a held-out recent-period check.

  5. 05

    Dashboard

    The Research Lab shows every run, chart and trade, and exports reports.

The public interactive demo is a separate static site: it replays synthetic demonstration data in the browser and does not run backtests on a server. The engine, data and strategies stay private. The demo source is public on GitHub.

04 / Development status

What works, what is next.

Working

  • Historical backtesting of crypto perpetual futures with fees, funding and slippage
  • Research Lab: run, quality check, compare runs, Markdown and JSON reports
  • Held-out recent 12-month check, counted separately from exploration
  • Strategy files with parameters and a fixed cost and risk policy
  • Public interactive demo (static, synthetic data)

Experimental

  • Exchange-sandbox demo runtime (private, no real funds)
  • Market context views, for information only

Planned

  • AI analysis of structured results: risk explanations and testable hypotheses
  • Broader out-of-sample validation
  • Public early access
  • Features for research teams
05 / Known limitations

Stated plainly.

  • The Research Lab runs privately and is not publicly available.
  • One run works at a time.
  • Backtesting covers crypto perpetual futures only.
  • Public screenshots and video use synthetic market data; results on real markets are not published.
  • The current chart labels crypto prices as VND and times as UTC+7; this is a known display issue.
  • There is no live trading.
06 / Availability

Built for research. Preparing for early access.

AI Backtest Lab is under active development. The research workspace is private and commercial pricing has not been announced. Public early access is a roadmap goal, not an available service.

Explore Interactive Demo Discuss Research Needs Research disclaimer