Quick definition
Treeova's internal platform credit system (₮) used to meter AI inference costs. Users receive Triobols through their subscription plan and spend them when AI agents run analysis or execute strategies.
Why a credit system instead of usage billing
Raw token-based billing is opaque to most traders — you don't want to think about prompt tokens before clicking 'analyze position'. Triobols translate variable LLM pricing into a single predictable unit, so a user sees "this analysis costs ~120 ₮" and can budget against their plan allowance.
Live peg, no hardcoded markup
The Triobol cost of any feature is computed by a database trigger from the underlying model price and a markup multiplier stored in the economic configuration table. When OpenRouter changes a model's price, every feature that uses that model is repriced automatically — no code deploy, no drift.
Settlement and audit
Triobol settlement is atomic. The user's canonical balance lives on the profile table; the UI cache on user_agent_credits is clamped by a database trigger so it can never display more credit than the user actually has. Every grant, spend, and refund flows through versioned RPCs with advisory locking.
Go deeper
The full technical methodology behind Triobols is documented in the Treeova whitepaper series.
Read the Triobols whitepaper →