Start Run
Start a new simulation run.
Request body
'backtest' runs historical data instantly; 'live' schedules real-time simulated paper trading; 'alpaca_paper'/'alpaca_live' use Alpaca broker
Date to simulate trading for (single day mode)
Start date for multi-day simulation
End date for multi-day simulation
Max concurrent day simulations for multi-day mode
Actual cash deposited in brokerage (per day for multi-day)
Intraday margin multiplier (e.g. 4x)
List of ticker symbols to trade
LLM provider(s) to use. Multiple providers run in parallel for comparison.
host:port for Qwen Baseline vLLM server (required if qwen-baseline selected)
host:port for Qwen Trading RL vLLM server (required if qwen-trading-rl selected)
host:port for DeepSeek vLLM server (required if deepseek selected)
ID of the BrokerAccount to use (required for alpaca_paper/alpaca_live modes)
Stop loss as % of initial cash. Liquidates all positions when breached. None or 0 to disable.
Enable Best of N mode - run LLM N times and judge to select best response
Number of candidate responses to generate in Best of N mode
Minutes between analysis steps
Market open time (first LLM decision step)
Session force-close cutoff (prompt mandate/countdown render from this)
Response
Successful Response
Map of LLM provider to run ID
Changes
No recorded changes to this endpoint across all 1 revision of this API.