Skip to content
Project // 005

Synapse
L3

01

The Objective

Designing a modular L3 rollup architecture specifically optimized for AI-agent-to-agent transactions, enabling autonomous economic activity between machine learning systems.

The protocol introduces a novel gas model that prices computation based on neural inference complexity rather than EVM opcodes.

02

Architecture

Rollup Core

Custom L3 Sequencer

Agent Protocol

A2A Transaction Format

Gas Model

Inference-weighted Pricing

03

Protocol Results

TPS Target10k
Gas Savings~92%
Agent Slots256
Finality<1s