Interactive walkthrough

See YorN in Action

Three real scenarios: agent creation, team coordination, and self-improvement.

Scenario: Create an NBA Halftime Agent

Watch as a user types a natural language prompt describing an NBA halftime comeback strategy. The AI generates an inspectable agent, validates it, and deploys it to paper trading for review.

Step 0 / 0
PENDING
Prompt Input
PENDING
AI Generation
PENDING
Validation
PENDING
Paper Deploy
YorN Dashboard
app.yorn.ai/agents/builder
Dashboard Agents Markets History
$10,000.00
PAPER
G
PAPER TRADING — Simulation Mode
P&L
$0.00
WIN RATE
TRADES
0
AGENTS
0
Press ▶ Play to begin
AGENT BUILDER
"Buy YES on NBA games where the home team trails by 10+ points at halftime. Historical comeback data shows these are underpriced below 35 cents."
ENTRY LOGIC
home_deficit >= 10 AND price <= 0.35
EXIT LOGIC
price > 0.60 OR game_end
POSITION SIZE
2% per trade
RISK LIMITS
5 max positions, 8% daily loss
HalftimeDeficit RUNNING
Markets: 14 NBA Positions: 1 Portfolio: $9,936
BUY 20x YES LAL-BOS-Q4 @ 32c
YorN Agent Builder
DEPLOYED
[00:00] INPUT Waiting for strategy description...
[00:02] INPUT > "Buy YES on NBA games where the home team trails by 10+ points at halftime. Historical comeback data shows these are underpriced below 35 cents."
[00:03] AGENT Parsing strategy description...
[00:04] AGENT Identified: sport=NBA, trigger=halftime_deficit, threshold=10pts, side=YES, max_price=0.35
[00:06] AGENT Generating Strategy class via Claude...
[00:12] BUILD HalftimeDeficit strategy generated and ready for review
[00:13] BUILD Entry logic: home_deficit >= 10 AND current_price <= 0.35
[00:13] BUILD Exit logic: price > 0.60 OR game_end
[00:13] BUILD Position size: 2% of portfolio per trade
[00:13] BUILD Risk: max 5 concurrent positions, 8% daily loss limit
[00:15] VALID Running syntax validation...
[00:16] VALID Syntax: PASS
[00:17] VALID Risk params: PASS (within safe bounds)
[00:18] VALID Exchange compatibility: Kalshi PASS
[00:20] DEPLOY HalftimeDeficit deployed to PAPER mode
[00:21] DEPLOY Scanning 14 active NBA markets...
[00:24] SIGNAL LAL-BOS-Q4: home deficit = 12pts, current price = 32c SIGNAL TRIGGERED
[00:25] TRADE BUY 20x YES @ 32c on LAL-BOS-Q4 (paper)
[00:25] STATUS Agent running. Portfolio: $9,936.00 | Positions: 1 | Scanning...

Scenario: Council Mode Vote on FOMC Trade

Four specialized agents analyze an FOMC rate decision market. Each agent evaluates the trade from its unique perspective, casts a weighted vote, and the council reaches consensus before execution.

Step 0 / 0
PENDING
Signal Detected
PENDING
Agent Voting
PENDING
Consensus
PENDING
Execution
YorN Dashboard
app.yorn.ai/agents/council
Dashboard Agents Markets History
$12,847.00
PAPER
G
PAPER TRADING — Simulation Mode
P&L
+$247.30
WIN RATE
68%
TRADES
34
AGENTS
4
Press ▶ Play to begin
COUNCIL VOTE — FOMC-MAR-HOLD @ 78c
MacroAlpha
YES
87% · 1.2x weight
SentimentBot
YES
72% · 1.0x weight
RiskGuard
YES
65% · 1.0x weight
ContrarianX
NO
54% · 0.8x weight
CONSENSUS REACHED
74.2% weighted YES (threshold: 60%) · Position: 0.8x
BUY 15x YES FOMC-MAR-HOLD @ 78c (paper)
MacroAlpha
YES
Confidence: 87%
Fed funds futures imply 92% hold probability. Current market at 78c is underpriced by 14 cents.
SentimentBot
YES
Confidence: 72%
FOMC statement language analysis suggests dovish hold. Social sentiment aligns with no-cut thesis.
RiskGuard
YES
Confidence: 65%
Risk/reward favorable at current price. Capped downside of 22c vs upside of 78c on hold outcome.
ContrarianX
NO
Confidence: 54%
Market pricing already reflects consensus. Limited edge at 78c. Would prefer entry below 70c.
CONSENSUS REACHED -- Execute Trade
3/4 agents voted YES · Weighted score: 74.2% (threshold: 60%) · Position size: 0.8x (adjusted for dissent)
Council Mode -- FOMC-MAR-HOLD
COUNCIL
[14:30:01] COUNCIL Signal detected: FOMC-MAR-HOLD @ 78c
[14:30:02] COUNCIL Initiating council vote with 4 agents...
[14:30:02] COUNCIL Consensus threshold: 60% weighted | Mode: simple majority
[14:30:04] VOTE MacroAlpha analyzing macro indicators...
[14:30:06] VOTE MacroAlpha votes YES (87% confidence, weight: 1.2x)
[14:30:06] REASON "Fed funds futures at 92% hold. Market underpriced by 14c."
[14:30:08] VOTE SentimentBot analyzing NLP sentiment...
[14:30:10] VOTE SentimentBot votes YES (72% confidence, weight: 1.0x)
[14:30:10] REASON "Statement language dovish. Social alignment: 81%."
[14:30:12] VOTE RiskGuard evaluating risk/reward...
[14:30:14] VOTE RiskGuard votes YES (65% confidence, weight: 1.0x)
[14:30:14] REASON "Favorable asymmetry. 22c risk vs 78c reward."
[14:30:16] VOTE ContrarianX running contrarian analysis...
[14:30:18] VOTE ContrarianX votes NO (54% confidence, weight: 0.8x)
[14:30:18] REASON "Consensus already priced. Limited edge at 78c."
[14:30:19] RESULT Weighted YES: 74.2% | Threshold: 60% | CONSENSUS REACHED
[14:30:19] ADJUST Position sized to 0.8x due to dissenting vote
[14:30:20] TRADE BUY 15x YES @ 78c on FOMC-MAR-HOLD (paper)
[14:30:20] AUDIT Full vote record logged to audit trail

Scenario: Self-Learning Improves Win Rate

Watch a PlayoffPsych agent go through three learning cycles. It analyzes its own losses, proposes strategy improvements, tests them on paper, and keeps only the changes that work. One cycle gets rolled back automatically.

Step 0 / 0
PENDING
Analyze Trades
PENDING
Propose Change
PENDING
Paper Test
PENDING
Evaluate
YorN Dashboard
app.yorn.ai/agents/learning
Dashboard Agents Markets History
$12,847.00
PAPER
G
PAPER TRADING — Simulation Mode
P&L
+$384.50
WIN RATE
52%
TRADES
47
AGENTS
2
Press ▶ Play to begin
PlayoffPsych SELF-LEARNING
Win Rate 52%
Sharpe Ratio 0.82
LEARNING HISTORY
CYCLE 1
52% --> 61%
rest-day filter
CYCLE 2
61% --> 48%
ROLLED BACK
CYCLE 3
61% --> 71%
momentum weight
Net: +19pp win rate Cycles: 3 Rollbacks: 1
CYCLE 1 -- KEPT
Win Rate 52% --> 61%
Sharpe 0.82 --> 1.24
Trades Analyzed 18
Added rest-day filter
CYCLE 2 -- ROLLED BACK
Win Rate 61% --> 48%
Sharpe 1.24 --> 0.61
Trades Analyzed 12
Tighter entry rejected
CYCLE 3 -- KEPT
Win Rate 61% --> 71%
Sharpe 1.24 --> 1.92
Trades Analyzed 24
Momentum weighting added
52%
Starting Win Rate
71%
Final Win Rate
3
Learning Cycles
1
Auto-Rollback
Self-Learning Engine -- PlayoffPsych
LEARNING
[CYCLE 1] ANALYZE Reviewing 18 trades from last 48 hours...
[CYCLE 1] ANALYZE Found pattern: 67% of losses occurred on back-to-back game days
[CYCLE 1] PROPOSE Add rest_days >= 1 filter to entry logic
[CYCLE 1] TEST Paper testing improvement for 12 trades...
[CYCLE 1] RESULT Win rate: 52% --> 61% (+9pp) | Sharpe: 0.82 --> 1.24
[CYCLE 1] KEEP Improvement kept. Strategy updated.
[CYCLE 2] ANALYZE Reviewing 12 trades since last cycle...
[CYCLE 2] ANALYZE Found pattern: larger deficits (15+) have higher reversion rate
[CYCLE 2] PROPOSE Tighten entry to deficit >= 15 (from >= 10)
[CYCLE 2] TEST Paper testing improvement for 8 trades...
[CYCLE 2] RESULT Win rate: 61% --> 48% (-13pp) | Sharpe: 1.24 --> 0.61
[CYCLE 2] ROLLBACK Performance degraded. Auto-reverting to previous version.
[CYCLE 2] RESTORE Strategy reverted to post-Cycle-1 state
[CYCLE 3] ANALYZE Reviewing 24 trades since last kept cycle...
[CYCLE 3] ANALYZE Found pattern: wins correlate with 3rd quarter scoring momentum
[CYCLE 3] PROPOSE Add Q3 momentum weighting to signal confidence
[CYCLE 3] TEST Paper testing improvement for 16 trades...
[CYCLE 3] RESULT Win rate: 61% --> 71% (+10pp) | Sharpe: 1.24 --> 1.92
[CYCLE 3] KEEP Improvement kept. Strategy updated.
[SUMMARY] COMPLETE 3 cycles complete | Net improvement: 52% --> 71% win rate | 1 rollback
[SUMMARY] STATUS PlayoffPsych resuming normal trading with updated strategy

What makes this different

Three capabilities that set YorN apart from every other trading tool.

Natural Language to Live Agent

Describe any strategy you can imagine. The AI generates a complete, deployable trading agent with entry logic, exit logic, position sizing, and risk rules. No code required at any step.

Team Intelligence

Agents collaborate like a professional trading desk. Council voting ensures high-conviction trades. Dissenting opinions automatically adjust position sizing for risk management.

Agents That Get Smarter

Self-learning cycles analyze performance, propose improvements, and automatically rollback changes that make things worse. Your agents improve with every batch of trades.

Ready to try it
yourself?

Everything you just saw is available right now. Build your first agent in minutes, not months.

No credit card required · Free paper trading forever