Intenus operates through a sophisticated four-layer architecture designed to deliver optimal trade execution.
The Journey of Your Trade
1
Intent Creation
When you express your trading goal, our Intent Agent transforms it into a standardized format called IGS (Intenus General Standard). This ensures your intent can be understood uniformly across the entire network.
Example: "Swap 1,000 SUI for at least 950 USDC within 5 minutes" becomes:
Intent Type: Swap (Exact Input)
Input: 1,000 SUI
Output: 950-1,000 USDC (range)
Max Slippage: 0.5%
Deadline: 5 minutes
2
Privacy Protection
Your intent is encrypted using Seal's programmable access control. This means:
Only qualified solvers can see your trade
Access is time-limited
Your strategy remains private from potential front-runners
3
Solver Competition
Multiple independent solvers (specialized trading algorithms) compete to propose the best execution path for your trade. Each solver:
Analyzes liquidity across all available protocols
Constructs optimal routing strategies
Submits their proposed solution with expected outcomes
4
AI-Powered Ranking
Here's where Intenus shines. Inside a secure Trusted Execution Environment (TEE):
AI models classify your intent type
Solutions are evaluated across multiple dimensions:
Profitability: How much output you'll receive
Reliability: Historical performance of the solver
Speed: Expected execution time
Risk: Slippage and failure probability
The best solution is selected through verifiable computation
5
Execution & Verification
The winning solution is executed on the Sui blockchain:
Transaction is atomically executed using Programmable Transaction Blocks (PTBs)
Results are verified against promised outcomes
Execution data is archived on Walrus for transparency
Poor-performing solvers face reputation penalties
Why This Approach is Superior
With Traditional DEX, you manually search for the best price across platforms. DEX Aggregators provide automated routing but limited to specific protocols. Intenus provides AI-powered competition across ALL protocols with verified, optimal execution