🇬🇧 English · 🇷🇺 Русский: README.ru.md

🛸 Cyber-Quant Signal Bot

Visual no-code builder for crypto trading strategies, backtesting & automated bots

Design algorithmic trading strategies by connecting blocks on an infinite canvas — indicators, Smart Money Concepts, order-flow, AI risk filters, and real execution — then backtest, optimize, and deploy them as standalone bots.

Features · Quick Start · Architecture · Roadmap · License

---

⚠️ Disclaimer — not financial advice. This software can place real orders on live exchanges. Trading crypto derivatives carries substantial risk of loss. Use paper trading first, never risk money you can't afford to lose, and review every strategy before going live. The authors accept no liability for financial losses.

---

✨ Features

Build a full trading pipeline — from data in to order out — visually, with 45+ node types:

• 📥 Data & exchanges — multi-exchange via CCXT (Binance, Bybit, OKX, MEXC, Kraken…), OHLCV + order-book streams, volume/volatility scanners, Polymarket whale tracking, Finviz (US stocks).
• 📈 Technical analysis — RSI, MACD, Bollinger, ATR, SMA/EMA, Stochastic, automatic divergence detection, multi-timeframe (MTF) filtering, math-expression nodes.
• 🏛 Smart Money Concepts — Fair Value Gaps, Order Blocks, BOS/CHoCH, Liquidity Sweeps, Daily Bias & Killzones.
• 📊 Order flow — Volume Delta / CVD, order-book wall distance, Deribit Put/Call ratio.
• 🧠 AI layer — Kronos time-series forecasting, an LLM risk filter (Hermes) that returns PASS/BLOCK with a confidence score, Local Deep Research for fundamental risk, and sentiment/ML filters.
• 💸 Execution & risk — Market/Limit orders, TWAP/VWAP, server-side OCO bracket orders, ATR-adaptive grid bots, Risk Guard (drawdown/exposure control), paper trading.
• 🔬 Test & optimize — full backtester, genetic parameter optimizer, visual execution-trail debugger.
• 🚀 Deploy — compile any strategy to a standalone Python + FastAPI bot in a Docker container; manage a fleet of bots with a Panic Stop.
• 🔌 Extras — PineScript v3–v6 importer (paste TradingView code → get a node graph, with quality report and coverage feedback), TradingView-style charts, bilingual UI (EN/RU), URL-based navigation (React Router), searchable node palette (140+ blocks).

🚀 Quick Start

Requirements: Docker Desktop. For the full AI stack you'll also want an NVIDIA GPU (Kronos) or an OpenAI-compatible LLM API (LM Studio, DeepSeek, etc.) for Hermes. The core platform runs without them.

git clone <your-repo-url> signal-bot
cd signal-bot
cp .env.example .env        # add your exchange / Telegram keys (optional to start)
docker compose up -d

Then open:

Service	URL
Web app	http://localhost
Backend API	http://localhost:3000/api
Kronos AI	http://localhost:8070
FreeLLMAPI	http://localhost:3456

💡 Start with paper trading and the built-in strategy templates — no API keys needed to explore the builder and backtester.

⚠️ Before exposing a server deployment: the compose file ships with development defaults — Postgres user/password and Local Deep Research admin/admin. Override POSTGRES_PASSWORD (and the matching DB_PASS), LDR_USERNAME/LDR_PASSWORD in your .env, and don't publish ports 5432/6450 beyond your private network.

🏗 Architecture

React + ReactFlow canvas  ─►  NestJS API (CCXT, signals engine, backtest, optimizer)
                                  │
                                  ├─ PostgreSQL  (strategies, signals, candles)
                                  ├─ Redis + Bull (cache, queues, pub/sub)
                                  ├─ Kronos        — time-series forecasting (GPU)
                                  ├─ Hermes        — LLM risk filter (OpenAI-compatible)
                                  ├─ FreeLLMAPI    — free-tier LLM aggregator (optional)
                                  └─ LDR + SearXNG — fundamental deep research

• Frontend: React 18, TypeScript, Vite, Zustand, React Router v7, ReactFlow 11, Lightweight Charts, CSS design tokens, ARIA accessibility.
• Backend: NestJS, TypeORM, PostgreSQL, Redis + Bull, CCXT — 26 modules (signals engine, backtest, optimizer, orders, paper-trading, risk, fleet, codegen, …).
• AI: Kronos (time-series forecasting), Hermes (LLM risk filter — supports OpenAI-compatible APIs: LM Studio, DeepSeek, Ollama), Local Deep Research, sentiment & ML filters.

🧭 Sitemap

Full navigation map: pages, modals, user flow — see SITEMAP.md.

Login / Signup
    │
    ▼
Dashboard ─────────────────────────────────┐
    │                                       │
    ▼                                       ▼
Builder (create strategy)          Strategies (browse / import)
    │                                  │
    ├─ Templates modal                 ├─ Edit → Builder
    ├─ Backtest modal                  ├─ Codegen → ZIP bot
    ├─ Optimizer (genetic)             └─ Backtest
    └─ Codegen (Python bot)
           │
           ▼
     Fleet (deploy & monitor)
           │
           ▼
     Paper Trading (forward-test)
           │
           ▼
     Signal History (journal)

14 pages · 5 sidebar groups · 12 modals · Real-time via WebSocket + polling

Group	Pages
Trading Hub	Dashboard · Cross-Exchange · Paper Trading · Signal History
Strategy Studio	Builder · Strategies · Backtest · Backtest Job
Intelligence Lab	ML Trainer
Fleet	Bot Fleet (Panic Stop)
Utilities	Documentation · Settings

🗺 Roadmap

• AST logic validation (Z3 SMT solver catches impossible conditions like RSI > 70 AND RSI < 30)
• PineScript v3–v6 importer with quality report and pine_block fallback nodes
• React Router navigation (shareable URLs, browser back/forward)
• Searchable node palette with category filtering (140+ blocks)
• ARIA accessibility & keyboard navigation
• Dynamic Top-50 relative-volume scanner (adaptive to market regime)
• Sandbox that test-runs generated Python bots before download
• Strategy marketplace — share strategies, earn rating, get subscription discounts
• Hosted cloud version (one-click, no setup)

🙏 Acknowledgements

This project integrates third-party open-source components and services:

AI & Research

• Kronos by NeoQuasar — time-series forecasting model. Architecture code is included with attribution (see kronos/NOTICE.md); weights are downloaded at runtime from Hugging Face.
• Hermes AI risk filter — built-in LLM-based risk filter node. Supports any OpenAI-compatible API (LM Studio, DeepSeek, Ollama, etc.) via environment variables. Configure HERMES_PROVIDER, HERMES_API_URL, HERMES_MODEL in .env.
• FreeLLMAPI — free-tier LLM aggregator proxy. Combines 16+ providers (Gemini, Groq, Mistral, etc.) behind one OpenAI-compatible endpoint. Included as an optional Docker service (port 3456).
• Local Deep Research — AI-powered fundamental research agent. Used as an external Docker image for the LDR node.
• SearXNG — privacy-first metasearch engine. Powers web search inside LDR research pipelines.

Exchange & Market Data

• CCXT — unified crypto exchange API library. Connects to 100+ exchanges (Binance, Bybit, OKX, MEXC, Kraken, etc.) for market data and order execution.
• Polymarket — prediction market. Used for whale-tracking and event-driven signals via the Polymarket node.
• Finviz — US stock screener. Powers the Finviz scanner node for equities data and news enrichment.
• Deribit — crypto derivatives exchange. Used for Put/Call ratio data in order-flow nodes.

Frontend

• React Flow — node-based canvas library for the visual strategy builder.
• Lightweight Charts by TradingView — financial charting library for TradingView-style price charts.

Infrastructure

• PostgreSQL — primary database for strategies, signals, candles, and user data.
• Redis + Bull — caching, message queues, and pub/sub for real-time updates and backtest job processing.

📄 License

Source-available, non-commercial. You may self-host and use this software for personal, non-commercial purposes. Commercial use — reselling, offering it as a paid service, or selling bots built on this code — requires a separate commercial license.

See LICENSE for the full terms. For commercial licensing, contact the author.

---

_{Built for traders who think in systems. ⚡ Not financial advice — trade responsibly.}