Blackglass
Full Source Code

Own a production-grade
algo trading platform.
Deploy it as yours.

8 microservices. 10+ ML trainers. React dashboard. 3 exchange integrations. 18 months of development. Acquire the complete codebase and run your own autonomous trading operation.

See Pricing
What's Included
8
Microservices
Python / FastAPI
200+
ML Features
58 modules, multi-timeframe
3
Exchanges
Alpaca, Coinbase, Kraken
18mo
Development
Production-tested

Build This Yourself: $330–490K

Or acquire the complete, working codebase today. Here's what it would cost to rebuild from scratch.

8 microservices (FastAPI + SQLAlchemy)
Cost / Time$80–120K4–6 months
ML pipeline + ensemble training (10+ trainers)
Cost / Time$100–150K6–9 months
Exchange integrations (Alpaca, Coinbase, Kraken)
Cost / Time$40–60K3–4 months
React dashboard + operations monitoring
Cost / Time$30–50K2–3 months
Infrastructure (Docker, AWS, TimescaleDB, Redis)
Cost / Time$20–30K1–2 months
Risk management (ATR stops, circuit breakers, drawdown)
Cost / Time$30–40K2–3 months
Walk-forward backtesting engine
Cost / Time$30–40K2–3 months
Total
$330–490K18–24 months

Built For

Fund Managers

Add an autonomous AI trading sleeve to your existing operation. Production-grade infrastructure that integrates with your team.

Prop Trading Firms

Stand up a new automated strategy desk quickly. Customize the ML pipeline to express your edge.

Fintech Entrepreneurs

Launch a trading product or robo-advisor. Skip 18 months of development. Go to market in weeks.

Independent Traders

Run institutional-grade infrastructure on your own terms. No subscriptions, no vendor lock-in. Your platform, yours to maintain.

Complete Inventory

Everything in the repository. Nothing held back.

Services

8 Python/FastAPI microservices, each independently deployable

Data Ingestion — Real-time WebSocket feeds from Alpaca, Coinbase, Kraken. Polygon.io historical backfill. Multi-timeframe bar aggregation (1m–1d). TimescaleDB storage.
Intelligence — Feature computation (200+ features, 58 modules), ensemble inference, signal generation, walk-forward backtesting engine.
Execution — Order routing to 3 exchanges, adaptive TP/SL management, risk controller, position tracking, full trade audit trail.
Learning — Model training pipeline with MAML meta-learning for few-shot adaptation to new market regimes.
Learning Bayesian — PyMC MCMC inference (JAX backend, 4-chain NUTS, 2000+ samples), multi-task learning, Optuna hyperparameter tuning.
Orchestrator — System bootstrap, workflow coordination, symbol group management, backtest orchestration.
Operations — Health monitoring, dashboards, alerting, reporting, analytics.
User Management — AWS Cognito auth, API key management, role-based access control.

ML Pipeline

Production trainers + 10 additional models ready to enable

Production Models — Gradient-boosting ensemble (LightGBM, XGBoost, CatBoost) plus MAML meta-learning, Multi-Task Learning, and Bayesian MCMC — multiple trainers, not a single model.
Ensemble Architecture — Channel A (directional core) + Channel B (context/uncertainty) + hard gates + 5D bucket arbitration.
Feature Engineering — 200+ features across 58 modules: technical indicators, liquidity, microstructure, order flow, sentiment, multi-timeframe.
Training Infrastructure — Triple-barrier labeling, Optuna optimization (100 trials), walk-forward validation, purged K-fold, Monte Carlo resampling.
Additional Trainers (Included) — Temporal Fusion Transformer, Reinforcement Learning (PPO/SAC), Mixture of Experts, Quantile Regression, Conformal Prediction — code present, ready to enable.

Exchanges & Data

3 exchanges integrated, adapter pattern for adding more

Alpaca — US equities — full order lifecycle, real-time quotes via WebSocket, paper + live modes. Production-tested.
Coinbase — Crypto — Advanced Trade API, JWT auth, WebSocket order updates, sandbox + production.
Kraken — Crypto — REST + WebSocket, rate limiting, order management.
Polygon.io — Historical OHLCV data backfill for training and backtesting.
Extensible — Adapter pattern — add new exchanges/brokers without modifying core logic.

Frontend & Monitoring

React dashboard + Next.js marketing website

Trading Dashboard — React 18, TypeScript, Vite, Tailwind CSS. Real-time P&L, positions, trade history. Bloomberg-inspired dark theme.
Model Visualization — Confidence scores, signal strength, ensemble breakdown per symbol.
System Monitoring — Service health, uptime, error rates, alerting dashboard.
Marketing Website — Next.js, Tailwind CSS, Stripe integration, contact forms, analytics (GA4 + Clarity).

Infrastructure

Docker, AWS, databases — production-ready deployment

Local Development — Docker Compose — full system running locally with one command.
AWS Deployment — ECR + SSM + EC2 scripts — single-command deploy per service. Terraform IaC.
Databases — PostgreSQL + TimescaleDB (time-series), Redis (pub/sub, caching, jobs), PgBouncer (connection pooling).
Migrations — Alembic migrations for every service. Schema versioned and reproducible.
Networking — Nginx reverse proxy, SSL termination, production configuration included.

Risk Management

Multi-layer protection — the system is designed to survive first

Position Sizing — Kelly Criterion variant adjusted by regime, volatility, and model confidence.
Stop-Loss / Take-Profit — ATR-based dynamic levels, configurable multipliers per direction, micro-SL for flat trades.
Daily Loss Limit — 5% max drawdown per day — automatically blocks new signals when exceeded.
Agent Drawdown — 10% rolling 7-day window per agent — automatic throttling on breach.
Circuit Breakers — Per-exchange (5 consecutive failures → halt) + SL cluster detection (3+ hits in 1hr → 24h lockout).
Architecture — Layered pipeline: Event → State → Feature → Cognition → Decision → Action → Risk → Feedback → Context → Audit → Governance. No layer can be bypassed.

Acquire the Platform

One-time purchase. Full source code. Own it forever.

Platform License
$75Kone-time

The complete codebase, documentation, and deployment support. Everything you need to run your own trading operation.

Full source code — all 8 services
ML pipeline + trained model configs
React dashboard
Docker + AWS deployment scripts
All database migrations
Full documentation
5 hours deployment support
1-hour architecture walkthrough call
30 days private Slack/Telegram
Book a Walkthrough
Recommended
Launch Package
$150Kone-time

Everything in Platform License, plus hands-on deployment, customization, and extended support to get you trading.

Everything in Platform License
20 hours deployment + customization support
3 architecture walkthrough sessions
Custom exchange or strategy integration (1)
Production deployment assistance
Performance tuning for your infrastructure
6 months private Slack/Telegram
Priority bug fixes during support period
Book a Walkthrough

Payment via wire transfer, crypto, or Stripe. Escrow available via Escrow.com.

Layered Trading Architecture

Every trading decision flows through every layer. No shortcuts. No bypasses. Built for survivability.

Event
State
Feature
Cognition
Decision
Action
Risk
Feedback
Context
Audit
Governance
Philosophy

Survivability over aggression. Capital preservation first. The system is designed to not lose money before it tries to make money.

System Modes

Research → Shadow → Probation → Live → Halted. Progressive trust system. No mode can be skipped.

Communication

Services communicate via HTTP REST and Redis pub/sub. Clean service boundaries. Each service is independently deployable.

Questions

The complete source code for the entire BlackGlass platform — all 11 microservices, ML pipeline, React dashboard, Docker deployment configs, database migrations, documentation. Everything in the git repository. You can inspect it, modify it, deploy it, and run it as your own.

Backend: Python 3.11+, FastAPI, SQLAlchemy 2.x, Pydantic 2.x, Poetry. ML: gradient-boosting trainers (LightGBM, XGBoost, CatBoost) running the production ensemble; PyMC + JAX (Bayesian MCMC), MAML, Multi-Task Learning, and Optuna available alongside. Frontend: React 18, TypeScript, Vite, Tailwind CSS. Infra: Docker, PostgreSQL + TimescaleDB, Redis, Terraform, AWS.

Yes. You own the source code. Add new exchanges, modify strategies, change the ML models, rebrand the dashboard — it is fully yours. The codebase is well-structured with clear service boundaries.

The Platform License includes 5 hours of deployment support and 30 days of private Slack/Telegram access with the builder. The Launch Package includes 20 hours of support, 6 months of access, and a custom exchange or strategy integration.

We do not promise specific returns. The platform has been actively developed and tested against US equities via Alpaca, and is currently in paper trading. Backtest and paper-trading results are not indicative of live performance and have inherent limitations. Profitability in live trading depends on market conditions, strategy parameters, capital, execution venue, and operator decisions. We provide the infrastructure and tools — no guarantees of any return outcome.

With Docker Compose, you can have the full system running locally in under an hour. Production deployment to AWS takes 1–2 days with the included infrastructure scripts.

Yes. One payment, you own the code forever. No recurring fees, no license keys, no vendor lock-in. Optional support packages available separately.

Skip 18 months of development.

Book a 15-minute walkthrough. See the codebase, the architecture, the dashboard. Then decide.

Book a Walkthrough

15 minutes. Screen share. No pressure.