THE RESEARCH COMES FIRST
Your edge starts with evidence.
One desk for market structure, competing theses and your paper record.
THE TRACK RECORD
Paper equity
No recorded equity points yet.
BEFORE THE NEXT IDEA
Challenge the thesis.
Find the assumption that matters. Write down what would prove it wrong.
C34 BUDGET BOT · OKX
Bot this signal
Pick a desk signal, set a USD budget, and start one paper position. Live stays greyed until both human gates are set.
Paper default. Live locked.
| botId | sym | side | budget | uPnL | state | age |
|---|---|---|---|---|---|---|
| none | ||||||
Start a research workflow
STRUCTURE, NOT CERTAINTY
Chart lab
Closed candles. Calculated levels. Scenarios with an exit from the thesis.
BTC / USDT
Select Refresh to request verified market candles.Begin with the data.
Load a timestamped candle set to inspect the chart and its scenarios.
No sample prices are shown.
FLICK THROUGH EVERY PERP
Perpetual chart browser
Bybit, Binance and OKX public perpetuals. Keyboard and swipe change the instrument, not the time axis. Device-local shortlist only.
Favorites, visited charts and notes stay in this browser. They are not synced. Open-on-exchange is a public market link and does not place an order.
Load a perpetual
Public venue candles. No invented prices.Browse every perpetual.
ArrowRight next, ArrowLeft previous. Swipe in Browse mode on a phone.
BETTER QUESTIONS. BETTER RESEARCH.
Research desk
17 workflows for your AI research. Open a prompt, add evidence, then copy or pin it.
MAKE YOUR THINKING VISIBLE
My research canvas
Keep the thesis, its evidence and the condition that changes your mind together.
PROCESS OVER PREDICTION
Paper journal
The latest recorded simulated fills. No curated winners and no invented track record.
Recent paper fills
ENGINE RECORDA fill is not a closed trade. The overview's closed-trade count comes from the audit validator. Fewer than 20 closed trades support observations, not established patterns.
HOW THIS DESK WORKS
Evidence before conviction.
Keep data, interpretation and a trading result distinct.
Source every observation
Chart lab uses public Blofin demo candles. It removes unfinished candles, rejects broken intervals, and labels stale data. It never substitutes generated prices.
Make the calculation visible
Structure compares the last close with 20- and 50-candle simple moving averages. Support and resistance use the previous 20 candles; ATR uses 14 true ranges. These are calculations, not AI forecasts.
Keep an opposing case
Every research prompt asks for sources, an opposing interpretation, invalidation, coverage gaps and the next check. A prepared prompt is not evidence of an executed research agent.
Measure the paper record
The existing audit gate uses profit factor ≥ 1.20, drawdown ≤ 15% and ≥ 40 closed trades. The 90-day observation period and both human capital approvals remain separate requirements. A research-gate pass does not authorize live trading.