Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/grab/cursor-talk-to-figma-mcp/llms.txt

Use this file to discover all available pages before exploring further.

​
Overview

The annotation_conversion_strategy prompt guides AI agents through converting manual annotations (numbered/alphabetical indicators with connected descriptions) into Figma’s native annotation system. This enables better collaboration, version control, and design documentation.

​
When to Use

Use this prompt when:
  • Converting numbered/lettered manual annotations to native annotations
  • Migrating legacy annotation systems
  • Automating annotation creation from design documentation
  • Standardizing annotation formats across teams
  • Cleaning up annotation layers after conversion

​
Process Overview

​
Step 1: Get Selection and Initial Setup

Get the frame or component containing annotations: 
// Get the selected frame/component
const selection = await get_selection();
const selectedNodeId = selection[0].id;

// Get available annotation categories
const annotationData = await get_annotations({
  nodeId: selectedNodeId,
  includeCategories: true
});
const categories = annotationData.categories;

​
Step 2: Scan Annotation Text Nodes

Identify annotation markers and descriptions: 
// Get all text nodes in the selection
const textNodes = await scan_text_nodes({
  nodeId: selectedNodeId
});

// Filter for annotation markers
// Markers typically have:
// - Short text content (single digit/letter)
// - Specific font styles (often bold)
// - Located in containers with "Marker" or "Dot" in name
// - Clear naming patterns ("1", "2", "3" or "A", "B", "C")

const markers = textNodes.filter(node => {
  const isShort = node.characters.length <= 2;
  const hasMarkerInPath = node.name.includes("Marker") || 
                          node.name.includes("Dot") ||
                          node.name.includes("Annotation");
  const isNumeric = /^\d+$/.test(node.characters);
  const isLetter = /^[A-Z]$/.test(node.characters);
  
  return isShort && hasMarkerInPath && (isNumeric || isLetter);
});

// Identify description nodes
// Usually longer text nodes near markers or with matching numbers
const descriptions = textNodes.filter(node => {
  return node.characters.length > 10 && !markers.includes(node);
});

​
Step 3: Scan Target UI Elements

Get all potential annotation targets: 
// Scan for all UI elements that could be annotation targets
const targetNodes = await scan_nodes_by_types({
  nodeId: selectedNodeId,
  types: [
    "COMPONENT",
    "INSTANCE",
    "FRAME"
  ]
});

​
Step 4: Match Annotations to Targets

Match each annotation to its target using multiple strategies:

​
1. Path-Based Matching (Priority 1)

function findTargetByPath(marker, targetNodes) {
  // Get marker's parent container name
  const parentName = getParentName(marker);
  
  // Remove annotation prefixes
  const cleanName = parentName
    .replace(/^Marker:\s*/i, '')
    .replace(/^Annotation:\s*/i, '')
    .trim();
  
  // Find UI elements with matching names in their path
  return targetNodes.find(target => 
    target.name.includes(cleanName) ||
    getNodePath(target).includes(cleanName)
  );
}

​
2. Name-Based Matching (Priority 2)

function findTargetByName(description, targetNodes) {
  // Extract key terms from description
  const keyTerms = extractKeyTerms(description.characters);
  
  // Look for UI elements whose names contain these terms
  return targetNodes.find(target => {
    const targetNameLower = target.name.toLowerCase();
    return keyTerms.some(term => 
      targetNameLower.includes(term.toLowerCase())
    );
  });
}

function extractKeyTerms(text: string): string[] {
  // Remove common words and extract meaningful terms
  const commonWords = ['the', 'a', 'an', 'and', 'or', 'but', 'in', 'on', 'at'];
  const words = text.toLowerCase().split(/\s+/);
  return words.filter(word => 
    word.length > 3 && !commonWords.includes(word)
  );
}

​
3. Proximity-Based Matching (Fallback)

function findTargetByProximity(marker, targetNodes) {
  // Calculate marker center point
  const markerCenter = {
    x: marker.absoluteBoundingBox.x + marker.absoluteBoundingBox.width / 2,
    y: marker.absoluteBoundingBox.y + marker.absoluteBoundingBox.height / 2
  };
  
  // Find closest UI element
  let closest = null;
  let minDistance = Infinity;
  
  for (const target of targetNodes) {
    const targetCenter = {
      x: target.bbox.x + target.bbox.width / 2,
      y: target.bbox.y + target.bbox.height / 2
    };
    
    const distance = Math.sqrt(
      Math.pow(targetCenter.x - markerCenter.x, 2) +
      Math.pow(targetCenter.y - markerCenter.y, 2)
    );
    
    if (distance < minDistance) {
      minDistance = distance;
      closest = target;
    }
  }
  
  return closest;
}

​
Step 5: Apply Native Annotations

Convert matched annotations using batch processing: 
// Prepare annotations array
const annotationsToApply = [];

for (const marker of markers) {
  // Match with description
  const description = findMatchingDescription(marker, descriptions);
  
  if (!description) continue;
  
  // Find target using multiple strategies
  const target = 
    findTargetByPath(marker, targetNodes) ||
    findTargetByName(description, targetNodes) ||
    findTargetByProximity(marker, targetNodes);
  
  if (target) {
    // Determine appropriate category
    const category = determineCategory(description.characters, categories);
    
    // Determine properties based on content and target type
    const properties = determineProperties(description.characters, target.type);
    
    annotationsToApply.push({
      nodeId: target.id,
      labelMarkdown: description.characters,
      categoryId: category.id,
      properties: properties
    });
  }
}

// Apply annotations in batches
if (annotationsToApply.length > 0) {
  await set_multiple_annotations({
    nodeId: selectedNodeId,
    annotations: annotationsToApply
  });
  
  console.log(`Applied ${annotationsToApply.length} native annotations`);
}

​
Helper Functions

​
Determine Category

function determineCategory(text: string, categories: any[]) {
  const textLower = text.toLowerCase();
  
  // Check for keywords that suggest category
  if (textLower.includes('interaction') || textLower.includes('click')) {
    return categories.find(c => c.name.includes('Interaction'));
  }
  if (textLower.includes('content') || textLower.includes('text')) {
    return categories.find(c => c.name.includes('Content'));
  }
  if (textLower.includes('style') || textLower.includes('color')) {
    return categories.find(c => c.name.includes('Style'));
  }
  
  // Default to first category or create general category
  return categories[0] || { id: 'general', name: 'General' };
}

​
Determine Properties

function determineProperties(text: string, targetType: string) {
  const properties = [];
  
  // Add properties based on content and target type
  if (targetType === 'INSTANCE' || targetType === 'COMPONENT') {
    properties.push({ type: 'component-annotation' });
  }
  
  if (text.includes('important') || text.includes('critical')) {
    properties.push({ type: 'high-priority' });
  }
  
  return properties;
}

​
Find Matching Description

function findMatchingDescription(marker: any, descriptions: any[]) {
  const markerValue = marker.characters;
  
  // Look for descriptions that start with the marker value
  return descriptions.find(desc => {
    const descText = desc.characters;
    return descText.startsWith(markerValue + '.') ||
           descText.startsWith(markerValue + ':') ||
           descText.startsWith(markerValue + ')') ||
           desc.name.includes(markerValue);
  });
}

​
Complete Example

async function convertAnnotations() {
  // Step 1: Setup
  const selection = await get_selection();
  const nodeId = selection[0].id;
  const annotationData = await get_annotations({ 
    nodeId, 
    includeCategories: true 
  });
  
  // Step 2: Scan text nodes
  const textNodes = await scan_text_nodes({ nodeId });
  const markers = textNodes.filter(isAnnotationMarker);
  const descriptions = textNodes.filter(isAnnotationDescription);
  
  // Step 3: Scan targets
  const targets = await scan_nodes_by_types({ 
    nodeId, 
    types: ["COMPONENT", "INSTANCE", "FRAME"] 
  });
  
  // Step 4: Match and prepare
  const annotations = markers.map(marker => {
    const description = findMatchingDescription(marker, descriptions);
    const target = findTargetByPath(marker, targets.matchingNodes) ||
                   findTargetByName(description, targets.matchingNodes) ||
                   findTargetByProximity(marker, targets.matchingNodes);
    
    if (target && description) {
      return {
        nodeId: target.id,
        labelMarkdown: description.characters,
        categoryId: determineCategory(description.characters, 
                                     annotationData.categories).id,
        properties: determineProperties(description.characters, target.type)
      };
    }
    return null;
  }).filter(Boolean);
  
  // Step 5: Apply
  await set_multiple_annotations({ nodeId, annotations });
  
  return `Converted ${annotations.length} annotations`;
}

​
Related Prompts

​
Related Tools

Build docs developers (and LLMs) love

Get started for freeTalk to us