Built for traders who value control
Aegis brings automation to crypto trading without asking you to hand over your funds. Your wallet is your identity, your keys stay yours, and every action is deterministic and auditable.
Most trading tools force a trade-off: give up custody and control to get automation. We do not accept that trade-off. Aegis never holds your funds — it connects to your exchange with trade-only, withdrawal-disabled API keys, or lets you sign every transaction yourself on a DEX. Strategies run on a deterministic quant engine with explicit risk limits, not opaque black boxes, and automation always starts disabled until you opt in. Our goal is simple: the leverage of automated execution with the safety of self-custody.
Principles that shape every decision
Four commitments guide how Aegis is engineered.
Non-custodial
Aegis never holds your funds. Exchange keys are rejected unless withdrawals are disabled, and on DEXs you sign every transaction yourself. The API never returns your secret.
Deterministic
Signals and risk decisions follow explicit rules, not black boxes. Every strategy uses defined stops and targets, and the risk engine behaves the same way every time.
Auditable
A double-entry ledger and an immutable audit trail record every move. Balances are derived from append-only sums, so you can always trace exactly what happened and why.
Wallet-first
Your wallet is your identity. Sign in with Ethereum — no email, no password, no sign-up. Link several wallets and recover your account from any verified one.
Purpose-built services, each for one job
Aegis is a set of focused services, each using the right language for the task.
Trading Core (Go)
The execution and orchestration layer, written in Go with Fiber, PostgreSQL, NATS JetStream and Redis. It owns orders, the ledger, the risk engine and the audit trail.
Quant Engine (Rust)
A dedicated Rust service exposed over gRPC. It computes the ten technical indicators, runs the five built-in strategies and hosts ONNX-model inference.
Training Pipeline (Python)
A Python pipeline that trains models with LightGBM and XGBoost and exports them to ONNX for the Quant Engine. Live on-chart inference is in development.
Dashboard (Next.js)
The interface you use, built with Next.js and React. It talks to the backend for wallet auth, strategies, backtests, orders and real-time updates.
These services communicate over gRPC and NATS JetStream, keeping execution, quant analysis and model training cleanly separated.
Control and automation, together
Connect your wallet and start in paper mode — no funds at risk.
Visit the Platform