Open JeongYunLee opened 4 months ago
Hello,
The content field in the search results provides the most query-related content from the scraped URL. Tavily uses proprietary AI and algorithms to extract only the most relevant content, optimizing for context quality and size rather than including the entire text from the source.
If you need the entire content from a source, you can utilize the Tavily Extract API, which is designed to retrieve the raw content from a list of websites. This is particularly useful for applications like data collection, content analysis, RAG and research.
To get started, refer to the documentation here: Tavily Extract API Documentation.
We also encourage you to join our Tavily Community to stay updated on the latest features, share feedback, and connect with other developers working with Tavily.
Hi, I'm building a langgraph using tavily and have a question.
When using tavily's basic search code in Python, I noticed that different sources of information are retrieved depending on the language of the query, even if the question is the same. When I use a language other than English (in my case, Korean), the length of the sentences in the content of the results is too short.
For example, in
example1
, when I ask "who is Messi?", the content length is quite long. However, inexample2
, when I ask the same question in Korean, the content part is shortened.My question is, is there a way to output the entire content without shortening it? I need the full content because I use the search results as the retrieve values for RAG.
Additionally, are the sentences in the content part the entire text from the URL, or is there an algorithm that retrieves only part of the text?
Looking forward to your response! Thanks!