What we have built so far
TTT is not released yet. This is what has been achieved during development: what got faster, what it can do now, and what you can rely on. No code, just results.
October 2026
- Opens in half a second: The first frame is on screen 0.33 to 0.5 seconds after launch, and a workspace of eight charts, each with its bars, is ready in 2.1 seconds. Performance
- Smooth with a lot on screen: A chart costs 2.4 to 4 milliseconds of work per frame. With 32 charts open the median is 4.6 milliseconds and memory stays near 210 MB. Going back to a timeframe you already visited takes 17 milliseconds. Performance
- Light on memory: One footprint chart sits at 360 to 390 MB; eight charts at about 250 MB; everything switched on at once (15 chart layers, ten indicators, ladder and tape) at 444 MB. Performance
- A VWAP suite built for order flow: Session VWAP that restarts at the open you choose (midnight UTC, London, New York), weekly, monthly, quarterly and yearly VWAPs together, anchored and fixed-range VWAPs you place on the chart, and a rolling VWAP. Every one carries its three standard-deviation bands, and a deviation panel with divergences shows how stretched price is. They are computed from the prices trades actually happened at, not from candles, and stretches built from candles alone are drawn fainter so an estimate is never mistaken for the real thing. Feature
- Closing the terminal never costs you work: Your workspace and your demo-account trades are saved every 20 seconds and again first thing when you close. Closing is quick, and a slow network can no longer make the window hang. Reliability
- Saved state survives a crash or a power cut: Layouts, drawings and demo accounts are written whole or not at all. If a file ever cannot be read, a copy of what it held is kept beside it instead of being replaced. Reliability
- Long history, kept in order: Captured market data is stored per symbol and per day under one budget. When a symbol reaches it, the oldest day goes first, across every kind of data at once, so a backtest never runs over a half-deleted history. Data
- Data you can trust is the right data: Spot, perpetual, inverse and options instruments are never mixed in one series, and a backtest states which markets it used. Where capture was interrupted, the gap is recorded and drawn on the chart instead of being filled in. Data
- T0 Lang: order-flow scripting with backtests you can prove: Write indicators and strategies that read the whole market: footprint cells, the book and its history, the tape, liquidations, open interest, funding, options gamma by strike, spot against perpetual, across venues. Monitors run continuously, and the same script runs live, in replay and in a backtest, with no look-ahead by construction. A backtest is deterministic: anyone who runs it again gets the same result, down to the hash. Feature
- Alerts that leave the terminal: A script can raise an alert as a notification, a sound or a voice, or send it to Discord, Telegram or your own address, with a chart thumbnail, under a permission you grant and a rate limit. Feature
September 2026 — the foundation
- Footprint, liquidity heatmap and depth ladder: Order flow at every price, a resting-liquidity heatmap you can scroll back through, and a depth ladder with the tape, in one GPU-rendered desktop window. Feature
- Derivatives context in the same place: Liquidations, funding, open interest, basis and options gamma, and order flow by counterparty on Hyperliquid, without leaving the chart. Feature
- A full charting workstation: 85 drawing tools, 61 indicators, 15 chart layers and 8 bar types, with a demo account to practise on stored data. Feature
- Held up under load: On a simulated venue the capture kept up with 1,600 trades a second on one instrument, 100 instruments at once, and 200 terminals on one backend, and recovered every time the venue dropped, went silent or repeated trades. Reliability
- More than 5,000 automated tests: Aggregation, indicators, drawings, the demo account, order construction, book reconstruction and replay determinism are covered by tests that run on every change. Reliability
Figures are measured on our own development laptop with release builds, and on a simulated venue where stated. They describe TTT only; we have not measured any other product.