Research by conversation
Ask the agent to investigate markets, companies, filings, strategies, and your connected account context.
Trading deployment guide
Use backtests for historical logic, paper trading for operational rehearsal, and live trading only after a separate review of broker permissions, execution risk, monitoring, and capital exposure.
Ask the agent to investigate markets, companies, filings, strategies, and your connected account context.
When direct trading is enabled, request a one-time trade in plain English, inspect the order, and approve it before submission.
Create, revise, backtest, connect, and operate Lumibot strategies without assembling the full application stack yourself.
Work in BotSpot or connect through ChatGPT, Claude, Cursor, Codex, and other compatible MCP clients.
Backtesting, paper trading, and live trading answer different questions. Treating them as interchangeable hides evidence gaps.
Paper trading can reveal scheduling mistakes, stale state, malformed orders, missing data, authentication problems, and monitoring gaps that historical tests miss.
Observe trade and no-trade periods, reconnects, restarts, rejected actions, and agreement between strategy state and broker state.
Paper orders are simulated rather than routed to a live market. Systems may omit or approximate market impact, information leakage, latency slippage, queue position, price improvement, fees, dividends, or available liquidity.
Paper results must remain labeled as simulated, never live performance.
BotSpot broker modes vary. Some connections support paper and live workflows; others are live-only. Authentication, assets, permissions, market data, fees, order eligibility, and regions also vary.
Read the current BotSpot broker page and official broker documentation before connecting.
Passing a paper test does not authorize live use. Review code, historical assumptions, paper observations, broker configuration, exposure constraints, and monitoring responsibility first.
Define acceptable exposure, order constraints, escalation steps, and a stop process before deployment.
After live deployment, compare real order timing, fills, fees, rejections, and position state with backtest and paper assumptions. Investigate differences instead of dismissing them as noise.
Return to paper validation after material changes to logic, broker integration, data, credentials, scheduling, or infrastructure.
Paper trading is useful for operational testing, but it does not predict live results. Live behavior can differ because of latency, slippage, liquidity, fees, partial fills, market impact, outages, and changing conditions.
No. Paper trading validates parts of the workflow, but fills and account behavior remain simulated. It cannot prove profitability or live execution quality.
No fixed duration proves readiness. Testing should cover normal schedule, trade and no-trade conditions, restarts, monitoring, and likely failure paths.
That depends on the broker. Some use separate keys, accounts, or endpoints. Check the current broker page and official documentation.
Sources verified 2026-08-17.
Past performance does not guarantee future results. Automated trading involves risk of loss.