What is search_depth here mean? tavily-python/tavily/tavily.py at master · tavily-ai/tavily-python · GitHub
Hello there,
The search depth parameter determines whether your search is basic or advanced.
Our Advanced Search includes post-processing to ensure you receive the most relevant sources, delivering more accurate and refined results for your queries.
Please note:
- A basic search costs 1 API Credit
- An advanced search costs 2 API Credits
More enhancements are on the way—stay tuned for even better search capabilities!
Best,
May
Hey May,
can you be a bit more specific about the value provided by the “pos-processing”? How are the results more accurate and refined? Is it using more source URLs? Are you using a smarter model to extract the content from each site?
Thanks
Kurt
Hi!
To clarify, we’re not using models to extract content. However, our advanced search feature is optimized for retrieving the most relevant sources and content snippets to your query, rather than providing generic snippets from each source.
There’s a new parameter, chunks_per_source , which specifies how many chunks per source you’d like to include in the content snippet. The default is 3, but we recommend adjusting it to between 1 and 3, depending on your use case.
We’ve recently released an improvement that enhances this process and increases the likelihood of retrieving the requested number of chunks per source when you include the include_raw_content=true parameter in your query. However, please note that including raw content may slightly increase latency.
I hope this helps clarify the value of the post-processing improvements!
Best,
May
1 Like