Content
@
https://opensea.io/collection/dev-21
0 reply
0 recast
2 reactions
rish
@rish
Big upgrade on the cast search API today - semantic search We're storing millions of vectors, each vector made within seconds of when a cast hits the network. When a semantic or hybrid search is performed, we get the vector for the search query, find the casts with vectors that are closest in vector space and insert in the API response. Up next: - algorithmic ranking of search results (right now its desc chron only) so the best results show at the top - any necessary latency improvements Feedback is useful as always πͺ https://docs.neynar.com/reference/search-casts
7 replies
4 recasts
47 reactions
Sam
@sams
Will vectors also be available through your indexing service?
1 reply
0 recast
0 reaction
rish
@rish
what is the use case you have in mind?
1 reply
0 recast
0 reaction
Sam
@sams
Say I'm fetching a list of casts for a feed or in this case a search. If I have some custom user metadata in another table in my db I would like to fetch along with each author of each cast (rank, xp level, points balance, etc...) its much more performant and cost effective to just make that relation in my db and get all the data I need in a single request
1 reply
0 recast
1 reaction
rish
@rish
will get back
1 reply
0 recast
1 reaction