Pine Graphics Deep Dive

The FATBot system reads indicator data through four Pine graphics MCP tools. These tools extract visual objects that Pine Script indicators draw on the chart — labels, lines, tables, and boxes.

The system never reads indicator plot() values. Everything comes through graphics objects. Indicators just need to be on the chart.

---

data_get_pine_labels

Reads Pine Script label.new() objects.

Default Return Format

{
  "labels": [
    { "text": "HUNT_LOW", "price": 5870.00 },
    { "text": "Bullish SMT", "price": 5868.50 }
  ]
}

Verbose Return Format (verbose: true)

{
  "labels": [
    {
      "id": "label_42",
      "text": "HUNT_LOW",
      "price": 5870.00,
      "x": 1713100800,
      "yloc": "price",
      "size": "normal",
      "textColor": "#00FF00",
      "color": "#00FF0033"
    }
  ]
}

study_filter Parameter

Limits results to labels from a specific indicator (substring match):

{ "study_filter": "Session Bias" }

---

data_get_pine_lines

Reads Pine Script line.new() objects. Returns horizontal price levels.

Default Return Format

{
  "horizontal_levels": [
    { "price": 5892.50 },
    { "price": 5870.00 }
  ]
}

Verbose Return Format (verbose: true)

{
  "horizontal_levels": [
    {
      "id": "line_15",
      "y1": 5892.50,
      "y2": 5892.50,
      "x1": 1713090000,
      "x2": 1713110000,
      "style": "solid",
      "width": 2,
      "color": "#FF5252"
    }
  ]
}

Line Styles

Line style field can encode meaning: - "solid" — active/unspent level - "dotted" — spent/mitigated level - "dashed" — pending/conditional level

---

data_get_pine_tables

Reads Pine Script table.new() objects. Tables are structured grids rendered in a chart corner.

Default Return Format

{
  "tables": [
    {
      "rows": [
        "Daily | Bullish | PDH",
        "Weekly | Bullish | --"
      ]
    }
  ]
}

Parsing

Split rows on | to extract cell values:

const cells = row.split(' | ');
// "Daily | Bullish | PDH" -> ["Daily", "Bullish", "PDH"]

---

data_get_pine_boxes

Reads Pine Script box.new() objects. Boxes represent rectangular price zones.

Default Return Format

{
  "zones": [
    { "high": 5878.00, "low": 5872.00 }
  ]
}

Verbose Return Format (verbose: true)

{
  "zones": [
    {
      "id": "box_8",
      "high": 5878.00,
      "low": 5872.00,
      "x1": 1713095000,
      "x2": 1713110000,
      "borderColor": "#4CAF50",
      "bgColor": "#4CAF5033"
    }
  ]
}

Color Encoding

Box colors can encode direction: - Green-family (borderColor green channel dominant): Bullish zone - Red-family (borderColor red channel dominant): Bearish zone

---

Common Parameters

All four tools share:

Parameter	Type	Default	Description
study_filter	string	none	Substring match on indicator name
verbose	boolean	false	Include IDs, colors, coordinates
max_labels	number	50	Cap on labels per study (labels only)

---

Usage Example

Read data from multiple indicators in parallel:

const [biasTable, sessionLabels, obBoxes, smtLabels] = await Promise.all([
  mcp.call('data_get_pine_tables', { study_filter: 'Bias' }),
  mcp.call('data_get_pine_labels', { study_filter: 'Session' }),
  mcp.call('data_get_pine_boxes',  { study_filter: 'OB', verbose: true }),
  mcp.call('data_get_pine_labels', { study_filter: 'SMT' }),
]);

Concurrent calls — total time bounded by slowest call (typically 200–500ms).