Recommendations

Time	Status	User Agent
Retrieving recent requests…

Overview

This endpoint processes free-text queries and returns contextually relevant location recommendations with supporting review citations. It combines Tripadvisor's vast database of locations with AI-powered search to understand user intent and deliver personalized results.

Key Features

Natural Language Processing: Understands complex, conversational queries
Review-Based Citations: Provides specific review snippets that justify each recommendation
Flexible Geography: Supports destination names, geo IDs, or precise coordinate-based searches
Category Filtering: Focus results on specific types of locations (restaurants, hotels, attractions, etc.)
Agentic-Friendly: Optimized response format for AI agent consumption and decision-making

Use Cases

robot

AI Travel Assistants

Build conversational travel bots that can understand queries like "romantic dinner spots in Paris" and provide specific restaurant recommendations with review-backed justifications.

route

Personalized Itinerary Builders

Create dynamic itinerary systems that adapt recommendations based on user preferences, travel style, and specific requirements mentioned in natural language.

building

Enterprise Travel Tools

Integrate into business travel platforms to provide curated recommendations for corporate travelers based on expense policies and preferences.

Request Examples

Simple Query by City Name

{
  "query": "best pizza places for a casual date night",
  "geo": {
    "name": "New York City"
  },
  "limit": 5
}

Advanced Query with Coordinates

{
  "query": "family-friendly attractions that are good for kids under 10",
  "geo": {
    "search_area": {
      "centroid_latitude": 40.7589,
      "centroid_longitude": -73.9851,
      "search_radius_meters": 5000
    }
  },
  "top_level_categories": ["Attraction"],
  "limit": 8,
  "response_preference": "quality"
}

Category-Filtered Search

{
  "query": "upscale restaurants with great wine selection",
  "geo": {
    "geo_id": 187147
  },
  "top_level_categories": ["Eat & Drink"],
  "limit": 3,
  "exclude_location_ids": [123456, 789012]
}

Error Handling

The API returns structured error responses with specific codes and messages:

Typical Error Responses

{
  "message": "Invalid request parameters",
  "type": "validation_error", 
  "code": "INVALID_PARAMETERS"
}

Best Practices for Agentic Workflows

Query Optimization

searchCrafting Effective Queries

Do:

Use natural, conversational language
Include context about the user's intent or preferences
Specify constraints like "family-friendly" or "budget-conscious"
Mention specific requirements or accessibility needs

Examples:

✅ "romantic restaurants with outdoor seating for anniversary dinner"
✅ "kid-friendly museums that are interactive and educational"
❌ "restaurants" (too vague)
❌ "food places" (lacks context)

Response Processing

quote-rightLeveraging Review Citations

The review_sources array provides valuable context for AI agents:

Quality Assessment: Use review snippets to understand why a location was recommended
User Matching: Match review content to user preferences mentioned in the query
Confidence Scoring: Locations with more relevant review citations may be better matches
Explanation Generation: Use snippets to explain recommendations to end users

Example Usage:

def explain_recommendation(result):
    location_name = result['location']['names'][0]['value']
    citations = result['review_sources']
    
    explanation = f"I recommend {location_name} because:"
    for citation in citations[:2]:  # Use top 2 citations
        explanation += f"\n- {citation['snippet']}"
    
    return explanation

Performance Optimization

Response Preference Settings:

Use "speed" for real-time chat applications
Use "quality" for detailed recommendation engines
Default "quality" provides best results for most use cases

Pagination Strategy:

Start with limit: 5 for initial results
Use exclude_location_ids to get additional results
Cache results to avoid duplicate API calls

Rate Limits

Standard Plans: 10 requests per second
Enterprise Plans: Custom rate limits available

Getting Started

Obtain API Key: Currently, the Agentic Search service is available as a per-request basis until a self-service sign-up is released in 2026 Q1. Contact your Tripadvisor representative for access.
Test with Simple Queries: Start with basic location searches to understand response format
Integrate Review Citations: Build logic to process and present review-based justifications
Optimize for Your Use Case: Adjust response_preference and filtering based on your application needs

Support

For technical support or questions about implementation:

Contact your Tripadvisor technical account manager
Use our developer support portal for API-specific questions