Cortex

really pumped for this launch!! @cortex, a perplexity-style chatbot for farcaster and onchain data, is now live as a v2 frame and online @ withcortex.com

ask questions about some of your favorite apps and protocols such as @farcaster, @icebreaker, @clanker, @wow, @eventsxyz, @bountycaster, and @streamm -- enhanced my rich widgets you can use to easily take actions

lmk if you have any questions/comments and check out the next cast for info re: open source contributions! more to come soon 🫡

A sufficiently decentralized social network. farcaster.xyz

Cortex is the command center for your online life, coming soon | withcortex.com

The open professional network where you meet exceptional people • Casting in /icebreaker • Get started at icebreaker.xyz

Bounties and services on Farcaster bountycaster.xyz. Tag @bountybot to post and check out /bounties. FAQ: bountycaster.xyz/faq

Streamm.tv - launch a meme token and go live!
/streamm

It’s FB Events for Farcaster. Follow @event for updates about your RSVPs. Built by @matthew.

create autonomous memecoins for free, get rewarded for posting bangers

building /cortex, hacking /farhack, yapping /dylan | dylansteck.com

Oh hell ya!! 🚢 
What are you using to index all that info?

Building Higher apparel at High Line 🧢 
Host at /clickcast /cpg || Husband to @shiwen

thanks! technically no indexing right now just a bunch of different tool calls at runtime

yeah literally zero RAG at all so far lol, implementing that is gonna make the quality of the answers a lot better when I do so

I’m also caching both on my API and with the prompts directly which has helped both with cost and speed

Wait so no RAG at all? Or at least not chunking that info somewhere in some form of vector store?