How This Helps

Semantic Search returns results based on the meaning of your query—not just keyword matches. It’s perfect for finding relevant content using natural language.

​
Overview

The GET /api/v1/explore/{dataset_id} endpoint enables semantic search over captions—using models that understand language and context, not just keywords.

Example query:

dog playing with a ball

May return:

  • “A golden retriever catching a frisbee in mid-air”
  • “Puppy chasing a red toy across the grass”
  • “Canine fetching a tennis ball at the park”

Results are sorted by relevance score (0.0–1.0).

GET /api/v1/explore/{dataset_id}

​
Query Parameters

  • image_caption (string): your semantic query
  • textual_similarity_threshold (float): e.g. 0.75
  • entity_type (string): e.g. IMAGES

​
Example (cURL)

curl -H "Authorization: Bearer YOUR_TOKEN" \
"https://app.visual-layer.com/api/v1/explore/95233006-eddc-11ef-b303-76dbc3993eb2?image_caption=dog%20playing%20with%20ball&textual_similarity_threshold=0.75&entity_type=IMAGES"

​
Sample Response

{
  "clusters": [
    {
      "id": "4e0e4d51-0fef-4fe1-a8ec-1a82b6f4880b",
      "entity_type": "IMAGES",
      "size": 3,
      "representative_id": "8dedd5e3-a95c-4d6c-aecb-bdf7c92b599a",
      "media": [
        {
          "id": "8dedd5e3-a95c-4d6c-aecb-bdf7c92b599a",
          "caption": "A dog playing with a red ball in the grass",
          "relevance_score": 0.93
        }
      ]
    }
  ],
  "total_pages": 1,
  "current_page": 0
}

​
Response Concepts

  • clusters: Groups of visually similar results
  • media: Items in each cluster
  • caption: AI-generated or user-supplied description
  • relevance_score: Semantic match score (0.0–1.0)

​
Advanced Filtering

You can add filters for labels and tags:

curl -H "Authorization: Bearer YOUR_TOKEN" \
"https://app.visual-layer.com/api/v1/explore/95233006-eddc-11ef-b303-76dbc3993eb2?image_caption=dog%20outdoor&textual_similarity_threshold=0.75&entity_type=IMAGES&labels=[dog,puppy]&tags=7b89e36c-c2d1-4af9-9d23-f74e018e67c5"

​
Retrieve Dataset Metadata

curl -H "Authorization: Bearer YOUR_TOKEN" \
"https://app.visual-layer.com/api/v1/explore/95233006-eddc-11ef-b303-76dbc3993eb2/exploration_metadata"
{
  "dataset_name": "Dog Images Dataset",
  "clusters_count": 42,
  "media_count": 1250,
  "labels": [
    { "name": "dog", "count": 870 },
    { "name": "person", "count": 350 }
  ]
}

​
Python Client Example

import requests
from typing import Dict, List, Optional

class VisualLayerClient:
    def __init__(self, api_url: str, api_token: str):
        self.api_url = api_url
        self.headers = {'Authorization': f'Bearer {api_token}'}

    def search_by_caption(self, dataset_id: str, caption_query: str, 
                          textual_similarity_threshold: float = 0.75,
                          entity_type: str = 'IMAGES',
                          page_number: int = 0,
                          labels: Optional[List[str]] = None,
                          tags: Optional[List[str]] = None) -> Dict:
        params = {
            'image_caption': caption_query,
            'textual_similarity_threshold': textual_similarity_threshold,
            'entity_type': entity_type,
            'page_number': page_number
        }
        if labels:
            params['labels'] = f"[{','.join(labels)}]"
        if tags:
            params['tags'] = "untagged" if tags == ["untagged"] else f"[{','.join(tags)}]"
        
        response = requests.get(f"{self.api_url}/api/v1/explore/{dataset_id}", headers=self.headers, params=params)
        response.raise_for_status()
        return response.json()
        
    def extract_media_info(self, results: Dict) -> List[Dict]:
        media_items = []
        for cluster in results.get('clusters', []):
            for item in cluster.get('media', []):
                media_items.append({
                    'id': item['id'],
                    'caption': item.get('caption'),
                    'relevance_score': item.get('relevance_score')
                })
        return media_items

​
Example Usage

from visual_layer_client_test import VisualLayerClient

# Initialize client with your API token
client = VisualLayerClient("https://app.visual-layer.com", "YOUR_TOKEN")

# Basic search
results = client.search_by_caption(
    dataset_id="2348e2ec-860c-11ef-968d-76f1445e4fca",
    caption_query="dog"
)

# Extract and print media results
media_items = client.extract_media_info(results)
print("Search results:")
for item in media_items:
    print(f"- ID: {item['id']}")
    print(f"  Caption: {item['caption']}")
    print(f"  Relevance: {item['relevance_score']}")
    print()

# Advanced search with filters
filtered_results = client.search_by_caption(
    dataset_id="2348e2ec-860c-11ef-968d-76f1445e4fca",
    caption_query="dog",
    textual_similarity_threshold=0.75,
    labels=["dog", "puppy"]
)

filtered_items = client.extract_media_info(filtered_results)
print(f"\nFound {len(filtered_items)} filtered results")