← Use cases
Market intelligence for AI agents
For agent builders, AI engineers, and quants shipping autonomous trading systems.
Autonomous agents make decisions under uncertainty. LLMs can describe markets but cannot price them. Agents need a calibrated probability surface they can query and pay for in real time.
Last reviewed May 2026 · 1 min read
How it works
Voyons is the probability primitive for autonomous agents.
- Calibrated scenarios per request. Each call returns thousands of plausible price paths for the chosen pair and horizon. Agents compute pre-trade probabilities, stress-test allocations, and evaluate counterfactuals against the full distribution.
- Pay-per-request via x402. No account, no API key. The agent signs a stablecoin payment and attaches it to the HTTP request. Open standard (HTTP 402, v1).
- Agent-native API. REST plus Python SDK. Structured numerical outputs, no natural language, no hallucinations.
In practice
A typical agent loop:
POST /v1/scenarios/generation/
{
"model": "best",
"n_scenarios": 4096,
"n_steps": 24,
"inputs": {
"asset": "BTCUSDT",
"timeframe": "1h"
}
}
The response carries 4,096 simulated BTC paths over 24 hours.
The agent computes P(drawdown > 5%) from the distribution and
rebalances only when conviction exceeds its threshold.
Full integration via the Duon Labs Python SDK. x402 flow at duonlabs.com/x402.