All case studies

Stratz.in

Live
Visit

Institutional-grade crypto trading engine for retail traders

Python
React
PostgreSQL
Redis
Exchange APIs
Docker

The Problem

Retail crypto traders had two options: manually execute every trade (slow, emotional, error-prone) or use a signal broadcaster in the cloud (unsafe, no per-account safety, no protective orders, no accounting). Neither option came close to what institutional trading desks use daily.

What We Built

Stratz.in is a governed multi-account execution platform. It is not a signal pusher. Every signal passes through a per-account safety and readiness pipeline before any exchange order is sent. The pipeline checks:

  • Account state and position alignment
  • Risk sizing and margin adequacy
  • Exchange-truth reconciliation
  • Market-data freshness gates
  • Idempotency and duplicate prevention
  • Protection-order readiness

Only after all checks pass does the system send an order to the exchange. Then it immediately places protective stop-loss and take-profit orders. It trails stops to breakeven. It adopts orphaned exposure. It reconciles accounting truth after every fill.

Architecture

The platform separates concerns into distinct execution layers:

  1. Signal evaluation — strategies publish trade signals with entry, stop-loss, and take-profit parameters.
  2. Per-account safety pipeline — each linked account is independently evaluated for readiness. The same signal may be valid for one account and rejected for another.
  3. Order execution — orders are sent via exchange API with idempotency keys, retry classification, and rate-limit handling.
  4. Protective order management — stop-loss and take-profit orders are placed immediately after entry fills. The system tracks breakeven trails, retries failed placements, and verifies each lifecycle step.
  5. Accounting convergence — fills, fees, and PnL are materialized into an accounting layer that converges to truth. Every trade can be reconstructed from signal decision through final status.

Production Metrics

These are real numbers from the production window (March–June 2026):

  • Zero broker-side 5xx or timeout create rejects
  • 0.14% rate-limit rejection rate (33 out of 23,843 orders)
  • Sub-second broker acknowledgement at p50: entry 391ms, TP 269ms, SL 170ms
  • ~9,000 take-profit placements, ~9,500 stop-loss trails
  • ~3,200 breakeven lifecycles with 643 retry events
  • 2,732 margin addition events
  • 856 orphan/open-exposure adoptions
  • 91 position reconstruction events

What This Proves

Stratz.in is live. It handles real money. It has executed tens of thousands of orders without a single broker-side failure. That's not a demo. That's production.

This is what Ootaboo builds: systems that don't break when it matters.

Engineering deep dive: How we built the institutional-grade crypto trading engine

Execution and reliability figures describe infrastructure performance, not trading returns, user profitability, or investment outcomes.

Production metrics

0
Strategy Orders Sent
0
Entry Orders Executed
0
Broker 5xx Rejects
0.14%
Rate-Limit Rejects
391ms
Entry Ack (p50)
~9,000 / ~9,500
TP/SL Placed
0
Orphan Adoptions
0
Breakeven Lifecycles

Technical model

Conceptual execution safety and acknowledgement path. This is a high-level technical diagram, not a view into proprietary strategy logic.

The accompanying Deep Dive explains the technical decisions and trade-offs behind this case study.

Read the trading-engine architecture deep dive