A fully autonomous multi-agent trading system with adversarial debate, macro intelligence, and self-learning. 5-dimension signal scoring, 6 hard risk controls, auto-submitted profit ladders, and post-trade AI debate — all server-side.
15 agents across 5 layers execute every cycle. Each feeds intelligence downstream in real-time. 3 always-on services run independently in the background.
Every position cycles through the full agent stack — no holding is ever blind
Position Sentinel scores each holding 0.0-1.0 across thesis, catalyst, and risk dimensions
Tier each position: WATCH (routine), CAUTION (trending weak), ALERT (thesis risk), CRITICAL (liquidate)
Full macro + sentinel + PM pipeline with thesis invalidation checks and news overlay
HOLD / TRIM / TIGHTEN_STOP / EXIT — each with execution timing (immediate, 1h, 4h, end of day)
Auto-submit stop orders + profit ladders to Alpaca. Concentration circuit breaker overrides thesis locks.
Every trade passes through 6 independent risk gates before execution. No agent can override these — they are enforced at the infrastructure level.
No single position exceeds 15% of total portfolio equity. Orders auto-reduced or blocked.
2+ positions or >25% weight in one sector blocks all new entries in that sector.
No new positions within 5 days of earnings date. Prevents overnight gap risk from binary events.
Consensus confidence must exceed institutional minimum. Learnable threshold — trade debates adjust it.
At least 3 of 4 analysts must be directionally aligned. Weak alignment triggers lower confidence.
Cash reserve scales by macro threat level: LOW 10%, MODERATE 15%, HIGH 20%, CRITICAL 30%.
Every closed position triggers a structured AI debate. The system learns from both wins and losses, automatically adjusting scanner weights, sector confidence, and conviction thresholds.
Defends entry timing, thesis quality, and sizing decisions. Identifies what WORKED with specific data points.
Attacks every decision — better entry price? More convergence needed? Should we have waited? Missed signals?
Weighs both sides, assigns verdict, and proposes concrete parameter adjustments for future trades.
These services run independently on their own schedules — feeding intelligence, enforcing risk rules, and learning from every trade.
AI researches geopolitics, Fed policy, sector rotation, supply chains. Generates threat-level digest that adjusts cash floor (10-30%) and injects sector-weight scoring modifiers into the scanner.
Position Sentinel scores each holding 0.0-1.0. Reeval PM issues HOLD/TRIM/EXIT with execution timing (immediate, 1h, 4h, end-of-day). Auto-submits stop orders to Alpaca.
3-round AI debate (Strategist vs Critic) analyzes each closed trade. Judge assigns WIN/LOSS verdict and proposes adjustments to sector confidence, scanner weights, and conviction thresholds.
When macro intelligence detects sector headwinds, the system doesn't wait — it cross-references your portfolio, adjusts scanner scoring weights, and scales the cash floor by threat level. Sector-only signals drive rebalancing.
AI + web search → geopolitics, central banks, sector rotation, supply chains
Threat level → cash floor adjustment + scanner weight modifiers per sector
"TIGHTEN_STOP UAL (airline headwinds) + boost scanner macroAlignment for #Defense sector"
Multi-agent LLM frameworks demonstrate consistent outperformance over single-agent baselines and buy-and-hold strategies across diverse market conditions in academic backtests.
View on OpenReview* Academic research results. Past performance does not guarantee future results.
Intelligence Scanner ranks 20+ candidates, then 4 parallel analysts, adversarial bull/bear debate with rebuttals, consensus judge, trade architect, risk analyst, PM verdict, and Alpaca execution.
Post-exit Strategist vs Critic with cross-examination. Judge assigns WIN_JUSTIFIED, WIN_LUCKY, LOSS_PREVENTABLE, or LOSS_UNAVOIDABLE. Feeds back into scanner weights and sector confidence.
Scanner scores across signalConvergence, fundamentalQuality, catalystProximity, technicalSetup, and macroAlignment. Produces both bullish AND bearish candidate rankings.
Profit ladder (3-5 GTC limit orders per position), hard and trailing stop orders auto-submitted to Alpaca. Reeval tightens/loosens stops automatically — no polling needed.
15% max position cap, sector concentration blocks, 5-day earnings blackout, 70% conviction gate (learnable), and macro-adaptive cash floor scaling 10-30% by threat level.
3-agent reeval pipeline: Macro Analyst + Position Sentinel (per-holding 0-1 scoring) + Reeval PM with execution timing — immediate, 1h, 4h, end-of-day, or next session.
Geopolitics, central banks, sector rotation analysis. Threat-level adaptive cash reserves (LOW 10% → CRITICAL 30%). Sector-weight scoring modifiers injected into scanner.
Every position records trader rationale, risk summary, PM approval note, conviction level (INSTITUTIONAL/CORE/TACTICAL), and horizon type (CORE_HOLD/MACRO_PLAY/OPPORTUNISTIC).
No hidden fees. Cancel anytime. All plans include paper trading — zero real money at risk.
Watch the agents think in real-time
Full AI trading desk, unlimited cycles
For serious quants and fund managers
For family offices, hedge funds, and high-net-worth individuals managing $1M+ portfolios. We deploy a fully private, dedicated instance of Desktop Wallstreet — tailored to your strategy, your risk profile, and your capital structure.
Private cloud deployment. Your data never touches shared servers. SOC 2 compliant isolation with 99.99% uptime SLA.
Agents tuned to your universe — equities, options, futures, crypto. Custom sector focus, position limits, and risk parameters.
Connect your prime brokerage — Interactive Brokers, Schwab, or institutional Alpaca. Multi-account execution with sub-account isolation.
Custom drawdown limits, overnight exposure caps, sector concentration ceilings, and real-time P&L guardrails — all configurable per account.
US equities, options chains, global ETFs, and macro instruments. Extend to crypto or commodities on request.
Dedicated solutions engineer. Strategy workshop, risk calibration, agent tuning, and ongoing optimization reviews — quarterly.
For portfolios $1M and above
Pricing based on portfolio size, number of accounts, and infrastructure requirements. Starts at $2,500/mo for single-family deployments.
Sign up in 30 seconds. No credit card. No risk. 15 agents start working immediately.