Trading-first crypto banking · live data, simulated execution

Trade like a fund.
Bank like the future.

OpeniBank fuses a pro-grade terminal with an AI banker that plans, simulates, and executes under receipt-backed controls — on rails built for real assets: iWallet custody, CreditChain settlement, the iUSD reserve.

150+
live markets
50×
max leverage
5
pillars, one bank
24/7
AI banker on desk

Markets, live now

Top pairs by volume — click through to the terminal.

All markets →

One bank, four surfaces

Trading first — everything else exists to make the trade safer, smarter, and settled.

The terminal

Trade

Hyperliquid-class execution surface: live depth, tape, leveraged paper engine, one-click order flow. Built dense, built fast, built for people who trade every day.

  • 150+ live markets
  • Order book + tape at 1.5s cadence
  • Up to 50× (simulated margin)
Launch Terminal
The edge

iBanker AI

A quant copilot on every chart — trend, momentum, book imbalance, taker flow, fused into bias and strategy. The full runtime plans, simulates, and executes under policy gates.

  • Live signal stack in the terminal
  • Plan → simulate → approve → execute
  • Receipts for every decision
Meet the agent
The yield layer

Earn

The OLP liquidity pool shares protocol PnL. iBanker strategy vaults put the same agent brain to work on your idle balance. The iUSD Prime Reserve is the cash leg.

  • OLP flagship pool
  • AI-managed strategy vaults
  • iUSD reserve at T-bill yield
Explore Vaults
The rails

Bank

Under everything: iWallet multi-chain custody, CreditChain settlement finality, and the OpeniBank ledger core enforcing permits, commitments, and receipts.

  • Multi-chain custody (EVM/BTC/SOL/TRON)
  • L1 settlement with BFT finality
  • Five-gate money boundary
See the stack

Product Architecture

Public web and desktop experiences above a shared banking core and an enterprise-grade agent runtime.

openibank/
├── openibank-web             ← Main public product for most users
├── ibanker-desktop           ← macOS-first rich financial client
├── ibanker                   ← Enterprise TypeScript agent runtime
├── shared-control-plane
│   ├── UAMS                  ← identity, sessions, entitlements, org roles
│   ├── account-plane         ← balances, wallets, subscriptions, accounts
│   ├── policy-plane          ← approvals, limits, delegated authority
│   └── evidence-plane        ← receipts, activity, audit, explanations
├── pillars
│   ├── iBanker               ← AI banker orchestration and tools
│   ├── iWallet               ← custody, wallets, recovery, balances
│   ├── CreditChain           ← settlement rails and programmable money
│   ├── iTrade                ← markets, routing, execution, portfolio motion
│   └── OpeniBank Core        ← banking OS spine and shared systems
└── agent-banking-subsystem
    ├── Ask                   ← understand customer or operator intent
    ├── Plan                  ← propose workflow and required approvals
    ├── Simulate              ← estimate outcome before live action
    ├── Approve               ← collect policy and human gates
    └── Execute               ← perform deterministic, receipt-backed actions

Non-Negotiable Invariants

Six rules that are never broken. They define OpeniBank's security model.

1

Identity, account authority, and policy are checked before privileged actions

2

Every value movement and approval path produces receipt-grade evidence

3

Simulation and production execution are clearly separated

4

Wallet custody and delegated authority stay explicit and auditable

5

AI banker actions are constrained by skills, tools, and approval boundaries

6

The same user and organization model spans web, desktop, mobile, and enterprise

The AI era of markets
starts at the terminal.

Live data, a $100k simulated book, and an AI banker on your desk — no signup, no keys, no risk. The stack is open-source under Apache 2.0.

terminal
$ git clone github.com/openibank/openibank
$ cd openibank
$ cargo test --workspace
test result: ok. 248 passed; 0 failed