Skip to content

marioxau/vault-trading-framework

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

IC SOLUTIONZ — Vault Trading Framework

Open-source multi-agent ICT/SMC signal framework for Deriv synthetic indices, Forex, and metals. Scans 32 pairs across multiple timeframes (1M → 4H), scores setups against your own strategy vault, debates them bull-vs-bear before approval, draws levels on TradingView, fires Discord alerts, and can auto-execute Deriv Multiplier contracts.

This repo is the framework only. It ships with an empty strategy vault (vault_template/) — you supply your own trading knowledge, Discord server, and broker credentials. No trading data, signals, or strategy content is included.

Risk disclaimer: trading synthetic indices and leveraged products carries a high risk of loss. This software is provided for educational and research purposes — nothing here is financial advice. Use demo mode first.


What it looks like

Automated chart markup — VOL75 SHORT setup with SSL sweep confirmed, inducement done, bearish OB entry, SL, TP1 (1:1R) and TP2 at the SSL target, all drawn by vault_tradingview_markup.py:

Auto-marked VOL75 SHORT setup

Educational cards generated by vault_image_generator.py and posted to Discord by the content scheduler:

Strategy library card Signal concept card
Strategy card Signal card

Architecture

TradingView Desktop (via tradingview-mcp, CDP)
        │
        ▼
ollama_prescan.py ──── local LLM pre-filter (H4+H1+M15 sweep check,
        │              FLAGGED / WATCH / SKIP tiers — token saver)
        ▼
vault_orchestrator.py ── full multi-TF scan per pair
        │
        ├── vault_brain_analyzer.py   confluence scoring vs your vault (0–4.0)
        ├── vault_debate_engine.py    bull vs bear analyst debate gate
        ├── vault_levels.py           key level detection
        ├── vault_tradingview_markup.py  Entry/SL/TP/OB/FVG drawn on chart
        │
        ▼
vault_discord_multi_post.py ── tiered Discord alerts (signals, risk log, digest)
        │
        ├── vault_price_alert.py     pm2 daemon — Deriv WebSocket tick monitor,
        │                            entry-tap + TP1/breakeven management alerts
        ├── vault_autotrader.py      auto-buys Deriv Multiplier contracts (demo/live)
        ├── vault_track_record.py    outcome eval — vault-true limit-fill model
        └── vault_journal.py         Obsidian/NotebookLM markdown journal export

Key design choices

  • Vault-driven scoring — the analyzer reads exact conditions from your .md strategy files, not hardcoded generic signals. Confluence ≥ 3.5/4.0 earns TIER_1_A_PLUS; only those are trade candidates.
  • Vault-true fill model — a setup only counts as filled when price physically reaches the limit/OTE zone. No phantom fills, no inflated win rates.
  • Debate gate — six analyst personas argue both sides; weak setups get rejected before they ever reach Discord.
  • Local-first — Ollama handles the high-volume prescan so cloud LLM tokens are spent only on setups worth analyzing.

Pairs covered

Group Symbols
Volatility VOL10/25/50/75/100 + 1s variants
Jump JUMP10/25/50/75/100
Crash/Boom CRASH500/1000, BOOM500/1000
Other synthetics STEP_INDEX, RDBULL, RDBEAR
Forex EURUSD, GBPUSD, USDJPY, AUDUSD, GBPJPY
Metals/Indices GOLD, SILVER, US30, NAS100, SPX500

Prerequisites

  • macOS or Linux, Python 3.10+, Node 18+
  • TradingView Desktop — works with any TradingView tier + tradingview-mcp (CDP bridge)
  • Claude Code — drives the scan protocol and analysis; works with Fable 5, Opus, and Sonnet models
  • A Discord server you control (bot token)
  • Ollama for the local prescan (optional but recommended)
  • Deriv account + API token (only for price alerts / auto-trading; demo supported)

Setup

git clone https://github.com/marioxau/vault-trading-framework.git
cd vault-trading-framework
pip3 install -r requirements.txt
npm install

# 1. Configure — interactive, writes .env (never committed)
node setup.js

# 2. Point the framework at your TradingView MCP clone
echo 'VAULT_MCP_PATH=/path/to/tradingview-mcp' >> .env

# 3. Build your vault — copy the template and write your strategies
cp -R vault_template ~/TRADING_KNOWLEDGE_VAULT
$EDITOR ~/TRADING_KNOWLEDGE_VAULT/01_STRATEGIES/

# 4. Run a scan
python3 vault_orchestrator.py

# 5. (optional) daemons — price alerts, scheduler
pm2 start ecosystem.config.cjs

See SETUP.md for the full Discord bot + webhook walkthrough and .env.example for every variable.

Repo layout

Path Purpose
vault_orchestrator.py Scan engine — all pairs, multi-TF, markup, Discord
vault_brain_analyzer.py Strategy-vault confluence scoring engine
vault_debate_engine.py Bull/bear debate gate
vault_price_alert.py Deriv WebSocket daemon — entry + management alerts
vault_autotrader.py Deriv Multiplier auto-execution (--demo default)
vault_track_record.py Outcome logging + vault-true evaluation
vault_journal.py Obsidian/NotebookLM journal export
ollama_prescan.py Local LLM prescan tiering
src/ Discord bot, webhook receiver, market feeds (Node)
vault_template/ Empty strategy vault — start here
examples/video_pipeline/ Optional YouTube content automation

License

MIT — see LICENSE.

Not affiliated with TradingView Inc. or Deriv. See the tradingview-mcp disclaimer for how chart access works.

About

Multi-agent ICT/SMC signal framework for Deriv synthetics, Forex & metals — vault-driven confluence scoring, bull/bear debate gate, TradingView markup, Discord alerts, Deriv auto-execution

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors