Travel Advisor Agent

Overview

The Travel Advisor agent is designed to assist users in making travel plans and is built entirely on the TaskingAI platform.

Common Usages

1. Recommend restaurants based on user preferences for food and budget.
2. Plan trip routes considering user preferences for time and views, as well as weather forecasts.
3. Analyze target locations (restaurants, hotels, spots) based on reviews.

Implementation

The Travel Advisor is constructed entirely on the TaskingAI platform without any coding required. In this section, we will demonstrate how to build a Travel Advisor and explain our choices for models and plugins. Here's an overview of the implementation:

‍

Model

Claude 3 Opus

The model for our Travel Advisor agent should have strong logical thinking ability, as it is expected to make the right plans given various variables, including user preference, time, weather, travel methods, etc. Here we choose the most powerful model in the Claude-3 family, the Claude 3 Opus model.

There are also some other models with good reasoning ability, for example, GPT-4, and Gemini 1.0 Ultra. You may choose one based on your preference.

‍

Memory

MessageWindowMemory

‍

System Prompt

"You are a global trip advisor knowledgeable about diverse travel destinations, local cuisines, accommodation options, and cultural practices worldwide. Your goal is to provide tailored travel advice based on the user's preferences, including ideal destinations, best times to visit, must-try foods, recommended accommodations, and essential cultural etiquette. Listen to the user's needs, ask clarifying questions if necessary, and offer insightful, practical travel planning tips that enhance their travel experience."

‍

Plugins

We equipped the Travel Advisor agent with four plugins: Trip Advisor, OpenWeather, DuckDuckGo, and Web Reader. Here's a closer look at each:

• Trip Advisor: A trip information platform that contains over 860 million reviews and opinions on nearly 8 million accommodations, restaurants, experiences, airlines, and cruises. This plugin provides the agent access to its extensive database to help users find travel information as requested.
• OpenWeather: An independent API provider that offers real-time weather information and forecasts. Weather information is crucial for planning trips as some activities depend on specific weather conditions.
• DuckDuckGo and Web Reader: This combination of plugins enables the agent to search for unfamiliar information and read content from the internet. DuckDuckGo is a widely used search engine with fewer restrictions than Google, although Google Search is also an available option on TaskingAI.

Integrate the agent into your own application

When you have a working agent with TaskingAI, it is simple to implement it in your own application. You may choose to use TaskingAI's Python Client SDK or REST API. Here is a simple example of connecting your agent using the Python SDK:

1import taskingai
2
3taskingai.init("YOUR_API_KEY")
4
5assistant = taskingai.assistant.get_assistant("YOUR_ASSISTANT_ID")
6
7# create a chat
8new_chat_session = taskingai.assistant.create_chat(
9    assistant_id=assistant.assistant_id,
10    name='new chat'
11)
12
13# send a user message
14user_message = taskingai.assistant.create_message(
15    assistant_id=travel_advisor.assistant_id,
16    chat_id=new_chat_session.chat_id,
17    text="Please help me and my partner plan a trip for 5 days to Italy. We are interested in art history, and our budget for food and hotel will be 1000 Euros per person exluding flights.",
18)
19
20# generate a user message
21assistant_response = taskingai.assistant.generate_message(
22    assistant_id=travel_advisor.assistant_id,
23    chat_id=new_chat_session.chat_id
24)

For further guidance on more complex implementations and to explore additional functionalities of TaskingAI, please refer to our comprehensive , please refer to the documentation.

‍