OmeSwap is a six-agent AI co-pilot for market decisions, risk control, strategy research, and chain-agnostic execution. Every signal comes with a clear Decision Receipt you can question at any depth.
OmeSwap is a chain-agnostic, agentic trading platform that combines a DeFi execution layer with a six-agent AI decision engine. It is designed as an explainable trading co-pilot where users can automate execution, get assisted recommendations, or run in research-only mode while keeping full visibility into why each decision is made.
A multi-agent ATS pipeline (Data, Signal, Graph, Regime, Risk, Execution) that collaborates on every trade decision.
A visual strategy builder (drag-and-drop DAG) with indicators, logic blocks, and backtesting workflows.
A DEX execution stack with AMM liquidity pools and multi-hop routing contracts.
A strategy/indicator marketplace flow with validation, publishing, and activation APIs.
Decision-receipt and memory infrastructure so actions remain auditable and explainable over time.
Explainability first: decisions are traceable to source signals, agent votes, and risk checks.
User-control by design: Autonomous, Assisted, and Solo modes support different trust and control levels.
Chain portability: execution adapters can be swapped while preserving core agent logic.
End-to-end ownership: from research and signal generation to risk-gated execution and reporting.
Frontend and product UX: Next.js + React terminal, portfolio, research, and builder interfaces.
Agentic intelligence layer: Python ATS services for orchestration, consensus, and risk-aware execution.
On-chain execution layer: Solidity contracts for liquidity pools, routing, and trade interactions.
Data and persistence layer: API routes, storage, and receipts for replayable decision history.
Traders who want AI-assisted execution with transparent reasoning.
Strategy creators who want to build, test, and publish reusable alpha logic.
Teams building chain-aware trading products that need a modular AI + DeFi foundation.
Last updated: May 16, 2026
During the hackathon, we evolved OmeSwap from a DEX prototype into an end-to-end agentic trading platform with explainable decisioning, on-chain execution, and a creator-focused strategy ecosystem.
Built the core 6-agent ATS pipeline:
Data Ingestion Agent
Regime Detection Agent
Signal Agent
Graph Agent
Risk Agent (veto authority)
Execution Agent
Implemented orchestration and auditability:
Multi-agent consensus flow in orchestrator
Decision Receipt generation for every execution cycle
API and conversation layer integration for explainable outputs
Shipped trading and strategy product surfaces:
Terminal/trade interfaces
Portfolio and research views
Visual strategy builder (node-based workflow)
Backtesting flow and creator publishing APIs
Delivered on-chain execution foundation:
Solidity contracts for multi-token pools
Multi-hop router support
Hardhat scripts for deploy, liquidity, token setup, and swap operations
Added infrastructure for marketplace and persistence:
Marketplace-related API routes and data models
Supabase migrations for strategies, activations, receipts, settings, and wallets
Phase 0 to Phase 8 implementation sequence is documented and completed in doc/phases/.
Frontend app, ATS backend services, and smart contract modules are now integrated in one repository.
Demo scripts (axl:agent, axl:demo) are in place for distributed ATS testing.
Coordinating multiple agents without tight coupling by using queue/state-driven flow.
Enforcing strict risk controls while preserving strategy flexibility for users.
Keeping decisions explainable by attaching structured receipts across the pipeline.
Unifying product UX (builder + terminal + marketplace) with execution infrastructure.
Working: End-to-end architecture, strategy lifecycle, and trading primitives.
In progress: Production hardening, deeper test coverage, and operational safeguards.
Next focus: Security reviews, reliability under load, and launch-grade monitoring.
OmeSwap now provides a practical template for chain-agnostic, explainable, AI-assisted trading: users can build strategies visually, validate them, activate them with guardrails, and understand every decision the system makes.
Currently bootstrapped by the founding team; no institutional round closed yet.
Not in an active priced round during hackathon development.
Preparing for an early fundraising process post-hackathon focused on security hardening, go-to-market, and multi-chain rollout.