Tavily Python Wrapper

The Tavily Python wrapper allows for easy interaction with the Tavily API, offering the full range of our search functionality directly from your Python programs. Easily integrate smart search capabilities into your applications, harnessing Tavily's powerful search features.

Installing

pip install tavily-python

Usage

Below are some code snippets that show you how to interact with our API. The different steps and components of this code are explained in more detail in the API Methods section further down.

Getting and printing the full Search API response

from tavily import TavilyClient

# Step 1. Instantiating your TavilyClient
tavily_client = TavilyClient(api_key="tvly-YOUR_API_KEY")

# Step 2. Executing a simple search query
response = tavily_client.search("Who is Leo Messi?")

# Step 3. That's it! You've done a Tavily Search!
print(result)

This is equivalent to directly querying our REST API.

Generating context for a RAG Application

from tavily import TavilyClient

# Step 1. Instantiating your TavilyClient
tavily_client = TavilyClient(api_key="tvly-YOUR_API_KEY")

# Step 2. Executing a context search query
context = tavily_client.get_search_context(query="What happened during the Burning Man floods?")

# Step 3. That's it! You now have a context string that you can feed directly into your RAG Application
print(context)

This is how you can generate precise and fact-based context for your RAG application in one line of code.

Getting a quick answer to a question

from tavily import TavilyClient

# Step 1. Instantiating your TavilyClient
tavily_client = TavilyClient(api_key="tvly-YOUR_API_KEY")

# Step 2. Executing a Q&A search query
answer = tavily_client.qna_search(query="Who is Leo Messi?")

# Step 3. That's it! Your question has been answered!
print(answer)

This is how you get accurate and concise answers to questions, in one line of code. Perfect for usage by LLMs!

API Methods

Client

NEW! We have released a Beta of our asynchronous Tavily client. It is available in version 0.3.4 of our Python package. The asynchronous client's interface is identical to the synchronous client's, the only difference being that all methods are asynchronous. Try it now with the AsyncTavilyClient class!

The TavilyClient class is the entry point to interacting with the Tavily API. Kickstart your journey by instantiating it with your API key. If you want to use Tavily asynchronously, you will need to instantiate an AsyncTavilyClient instead.

Once you do so, you're ready to search the Web in one line of code! All you need is to pass a str as a query to one of our methods (detailed below) and you'll start searching!

Methods

• search(query, **kwargs)

  • Performs a Tavily Search query and returns the response as a well-structured dict.
  • Additional parameters can be provided as keyword arguments (detailed below). The keyword arguments supported by this method are: search_depth, topic, max_results, include_domains, exclude_domains, include_answer, include_raw_content, include_images, use_cache.
  • Returns a dict with all related response fields. If you decide to use the asynchronous client, returns a coroutine resolving to that dict. The details of the exact response format are given in the Search Responses section further down.

• get_search_context(query, **kwargs)

  • Performs a Tavily Search query and returns a str of content and sources within the provided token limit. It's useful for getting only related content from retrieved websites without having to deal with context extraction and token management.
  • The core parameter for this function is max_tokens, an int. It defaults to 4000. It is provided as a keyword argument.
  • Additional parameters can be provided as keyword arguments (detailed below). The keyword arguments supported by this method are: search_depth, topic, max_results, include_domains, exclude_domains, use_cache.
  • Returns a str containing the content and sources of the results. If you decide to use the asynchronous client, returns a coroutine resolving to that str.

• qna_search(query, **kwargs)

  • Performs a search and returns a string containing an answer to the original query. This is optimal to be used as a tool for AI agents.
  • Additional parameters can be provided as keyword arguments (detailed below). The keyword arguments supported by this method are: search_depth (defaults to "advanced"), topic, max_results, include_domains, exclude_domains, use_cache,
  • Returns a str containing a short answer to the search query. If you decide to use the asynchronous client, returns a coroutine resolving to that str.

Keyword Arguments (optional)

• search_depth: str - The depth of the search. It can be "basic" or "advanced". Default is "basic" unless specified otherwise in a given method.

• topic: str - The category of the search. This will determine which of our agents will be used for the search. Currently, only "general" and "news" are supported. Default is "general".

• max_results: int - The maximum number of search results to return. Default is 5.

• include_images: bool - Include a list of query-related images in the response. Default is False.

• include_answer: bool - Include a short answer to original query. Default is False.

• include_raw_content: bool - Include the cleaned and parsed HTML content of each search result. Default is False.

• include_domains: list[str] - A list of domains to specifically include in the search results. Default is None, which includes all domains. Please note that this feature is only available when using the "general" search topic.

• exclude_domains: list[str] - A list of domains to specifically exclude from the search results. Default is None, which doesn't exclude any domains. Please note that this feature is only available when using the "general" search topic.

• use_cache: bool - Use the cached web search results. Default is True. If False is passed, a new web search will be done before generating your search results.

Search Responses

• answer: str- The answer to your search query. This will be None unless include_answer is set to True.

• query: str - Your search query.

• response_time: float - Your search result response time.

• images: list[str] - A list of query-related image URLs.

• results: list - A list of sorted search results ranked by relevancy. Each result is in the following format:

  • title: str - The title of the search result URL.
  • url: str - The URL of the search result.
  • content: str - The most query related content from the scraped URL. We use proprietary AI and algorithms to extract only the most relevant content from each URL, to optimize for context quality and size.
  • raw_content: str - The parsed and cleaned HTML of the site. For now includes parsed text only. Please note that this will be None unless include_raw_content is set to True.
  • score: float - The relevance score of the search result.
  • published_date: str (optional) - The publication date of the source. This is only available if you are using "news" as your search topic.

When you send a search query, the response dict you receive will be in the following format:

response = {
  "query" = "The query provided in the request",
  "answer" = "A short answer to the query", # This will be None if include_answer is set to False in the request
  "follow_up_questions": None, # This feature is still in development
  "images" = [
    "Image 1 URL",
    "Image 2 URL",
    "Image 3 URL",
    "Image 4 URL",
    "Image 5 URL"
  ], # This will be an empty list if include_images is not set to True
  "results" = [
    {
      "title": "Source 1 Title",
      "url": "Source 1 URL",
      "content": "Source 1 Content",
      "score": 0.99 # This is the "relevancy" score of the source. It ranges from 0 to 1.
    },
    {
      "title": "Source 2 Title",
      "url": "Source 2 URL",
      "content": "Source 2 Content",
      "score": 0.97
    },
  ] # This list will have max_results elements
}

Error Handling

The Tavily Python SDK includes comprehensive error handling to ensure smooth interaction with the API. Below are the specific exceptions that might be raised during usage:

1. Missing API Key: If no API key is provided when initializing the TavilyClient, a tavily.MissingAPIKeyError will be raised. Ensure you pass a valid API key to the TavilyClient during instantiation.

   from tavily import TavilyClient, MissingAPIKeyError
   
   try:
       tavily_client = TavilyClient(api_key="")
   except MissingAPIKeyError:
       print("API key is missing. Please provide a valid API key.")

2. Invalid API Key: If the API key provided is invalid, a tavily.InvalidAPIKeyError will be raised when sending a search query. Double-check that your API key is correct and active.

   from tavily import TavilyClient, InvalidAPIKeyError
   
   tavily_client = TavilyClient(api_key="invalid-api-key")
   
   try:
       response = tavily_client.search("Who is Leo Messi?")
   except InvalidAPIKeyError:
       print("Invalid API key provided. Please check your API key.")

3. Usage Limit Exceeded: If the API key provided is valid but the request fails due to exceeding the rate limit, surpassing the plan's monthly limit, or hitting the key's pre-set monthly limit, a tavily.UsageLimitExceededError will be raised. Consider upgrading your plan or checking your usage limits.

   from tavily import TavilyClient, UsageLimitExceededError
   
   tavily_client = TavilyClient(api_key="valid-api-key")
   
   try:
       response = tavily_client.search("Who is Leo Messi?")
   except UsageLimitExceededError:
       print("Usage limit exceeded. Please check your plan's usage limits or consider upgrading.")

These errors ensure that you are aware of the specific issues related to your API key usage, allowing you to take appropriate actions to resolve them.

License

This project is licensed under the terms of the MIT license.

Contact

If you are encountering issues while using Tavily, please email us at support@tavily.com. We'll be happy to help you.

If you want to stay updated on the latest Tavily news and releases, head to our Developer Community to learn more!