Skip to content

Data Flow

End-to-end walkthrough of how a TradingView chart becomes structured trading signals displayed on a live dashboard. Uses a concrete example: an ES 5-minute chart during the New York AM session.

Full Platform Flow

graph TB
    TV["TradingView Desktop<br/>ES 5min chart, 4 indicators"]
    TV -- "CDP WebSocket<br/>localhost:9222" --> MCP["tradingview-mcp<br/>data_get_pine_labels/lines/boxes/tables"]
    MCP -- "study_filter per indicator" --> READ["Readers<br/>5 parallel reads"]
    READ -- "normalized raw objects" --> INTERP["Interpreters<br/>contract rules applied"]
    INTERP -- "5 typed signals" --> ENG["Engine<br/>confluence + filters"]
    ENG -- "StrategyFlags JSON" --> SERVER["server.js<br/>localhost:8000"]
    SERVER -- "/api/data" --> BROWSER["Browser<br/>dashboard UI"]
    SERVER -- "nginx proxy" --> DOMAIN["ictedgefund.com"]

Scenario

  • Symbol: ES (E-mini S&P 500 futures)
  • Timeframe: 5 minutes
  • Time: 2026-04-14 09:55 ET (NY AM killzone, 0950-1010 macro active)
  • Chart state: Daily bias bearish, active macro window, bearish OB with FVG filter, SMT divergence between ES/NQ

Stage 1: TradingView Renders Pine Graphics

Four indicators run inside TradingView and render their output as Pine graphics objects. These exist in the DOM as structured data, not just pixels:

Indicator Renders
Bias AI Labels (bias direction) + tables (weekly bias, confidence)
Session Hunt AI Labels (session/macro IDs) + lines (liquidity levels)
OB AI Boxes (order block zones) + labels (FVG/mitigation status)
SMT AI Labels (divergence type) + lines (divergence connections)

Stage 2: MCP Reads via CDP

Each reader calls tradingview-mcp with study_filter to isolate its indicator. All five reads execute in parallel.

Bias AI labels (data_get_pine_labels({ study_filter: "Bias AI" })):

[
  { "time": 1744617600, "price": 5248.75, "text": "Bearish", "color": "#FF5252",
    "style": "label_down", "tooltip": "Daily Bias: Bearish | Confidence: 85%" },
  { "time": 1744185600, "price": 5260.00, "text": "Bearish", "color": "#FF5252",
    "style": "label_down", "tooltip": "Weekly Bias: Bearish" }
]

OB AI boxes (data_get_pine_boxes({ study_filter: "OB AI" })):

[
  { "left": 1744625400, "top": 5245.50, "right": 1744632600, "bottom": 5243.00,
    "backgroundColor": "rgba(255, 82, 82, 0.2)", "borderColor": "#FF5252",
    "text": "OB" }
]

Session Hunt AI labels + lines, SMT AI labels -- similar structured objects with timestamps, coordinates, colors, and text.

Stage 3: Readers Normalize

Each reader converts raw MCP responses into consistent shapes:

  • Unix timestamps to ISO strings
  • Coordinates into { time, price } objects
  • Related objects grouped (box + nearby label)
  • Out-of-range graphics filtered

No interpretation. The reader does not know what a red box means.

Bias reader output:

{
  "labels": [
    { "time": "2026-04-14T08:00:00Z", "price": 5248.75, "text": "Bearish",
      "color": "#FF5252", "tooltip": "Daily Bias: Bearish | Confidence: 85%" }
  ],
  "tables": []
}

Stage 4: Interpreters Extract Signals

Each interpreter applies its indicator contract to produce a typed signal.

{
  "type": "BiasSignal",
  "dailyBias": "bearish",
  "weeklyBias": "bearish",
  "confidence": 0.85,
  "biasShift": false
}
Rules: #FF5252 = bearish. Tooltip regex Confidence: (\d+)% extracts 0.85.

{
  "type": "SessionSignal",
  "activeSession": "NY_AM",
  "activeMacro": "0950-1010",
  "killzoneActive": true,
  "liquiditySweep": { "price": 5248.50, "direction": "above" }
}
Rules: Label text parsed for session + macro. Dotted line = liquidity level.

{
  "type": "OBSignal",
  "blocks": [{ "direction": "bearish", "high": 5245.50, "low": 5243.00,
                "fvgFiltered": true, "mitigated": false }],
  "nearestBlock": { "direction": "bearish", "distanceFromPrice": 0.75 }
}
Rules: #FF5252 border = bearish. Nearby "OB" label = active (not mitigated).

{
  "type": "SMTSignal",
  "divergencePresent": true,
  "direction": "bearish",
  "pair": "ES/NQ",
  "strength": 0.7
}
Rules: Label text "SMT Bear" + tooltip "ES/NQ Bearish Divergence".

{
  "type": "PriceSignal",
  "currentPrice": 5246.25,
  "atr": 12.5,
  "microStructure": "lower_highs"
}
Rules: ATR computed from OHLCV bars. Swing analysis on recent highs/lows.

Stage 5: Engine Scores and Builds Flags

No-Trade Filter Check

Filter Result
Daily bias established Pass (bearish)
Inside killzone Pass (NY AM)
Unmitigated OB visible Pass (bearish OB)
No bias shift Pass

All filters pass.

Confluence Scoring

Component Present Weight Score
Daily bias Yes 2 2
Weekly agrees Yes 1 1
Active killzone Yes 2 2
Active macro Yes 1 1
Unmitigated OB in bias dir Yes 3 3
FVG-filtered OB Yes 1 1
SMT in bias dir Yes 2 2
Total 12/12 100

Stage 6: StrategyFlags Output

{
  "timestamp": "2026-04-14T13:55:00Z",
  "symbol": "ES",
  "timeframe": "5",
  "bias": { "daily": "bearish", "weekly": "bearish", "shift": false, "confidence": 0.85 },
  "session": { "name": "NY_AM", "macro": "0950-1010", "killzone": true },
  "orderBlocks": [{ "direction": "bearish", "high": 5245.50, "low": 5243.00,
                     "fvg": true, "mitigated": false, "distanceFromPrice": 0.75 }],
  "smt": { "present": true, "direction": "bearish", "pair": "ES/NQ", "strength": 0.7 },
  "price": { "current": 5246.25, "atr": 12.5, "microStructure": "lower_highs" },
  "confluence": 100,
  "tradeReady": true,
  "setup": "SMT_OB_SESSION",
  "noTradeReasons": []
}

Stage 7: Dashboard Display

server.js serves this JSON at /api/data. The browser polls and renders:

  • Confluence gauge at 100 (green)
  • "TRADE READY" indicator lit
  • Setup type: SMT_OB_SESSION
  • Bearish bias badge with 85% confidence
  • NY AM session active, 0950-1010 macro highlighted
  • Order block zone drawn on chart screenshot
  • SMT divergence indicator for ES/NQ

nginx proxies localhost:8000 to ictedgefund.com with SSL termination via certbot.

Transformation Summary

Pine graphics (DOM objects)
    ↓  CDP Runtime.evaluate()
Raw JSON arrays (labels, lines, boxes, tables)
    ↓  study_filter isolation
Per-indicator raw data
    ↓  timestamp normalization, coordinate extraction
Normalized reader output
    ↓  color mapping, text parsing, spatial logic
Typed signals (BiasSignal, SessionSignal, OBSignal, SMTSignal, PriceSignal)
    ↓  no-trade filters, weighted scoring, setup detection
StrategyFlags JSON
    ↓  HTTP /api/data
Dashboard UI
    ↓  nginx reverse proxy + SSL
ictedgefund.com

Each transformation is a pure function of its inputs. No hidden state, no side effects, no feedback loops.