top | item 43248359

Show HN: Search and analyze millions of SEC filings with AI.

41 points| ashr_ | 1 year ago |publicview.ai

17 comments

order

dagelf|1 year ago

Bad UX... "chat with thousands of filings" but right after: pick a ticker to start. Sure, click the Chat tab instead. Type a query. Roadblock: you have to log in first. After login, boom query wiped. I'm sure people won't use it for the UX, but UX matters...

dagelf|1 year ago

Chat with "millions" of filings. As long as you select them one at a time.

ashr_|1 year ago

Good feedback, will work on saving the user's query before the signing in.

colesantiago|1 year ago

I've seen hundreds of these types of apps, to the point that I used Claude to build my own similar version of this app with similar features 100% for free.

What is the moat / advantage here when I or anyone else can build my own and at my own will serve it as a SaaS if I want to?

I'm assuming data isn't the moat here either as we all get the data from SEC Edgar.

pranavm27|1 year ago

From my perspective, the moat in such cases would be how many users impressed (based on accuracy, specific pain point) and how fast were they impressed.

Part of "how fast" equation is the UX - build with zero friction - every second they gonna churn unless impressed.

ashr_|1 year ago

The advantage here is the ease of searching for a filing. It's more convenient than having to search the EDGAR database then upload the doc to Claude every time.

janmo|1 year ago

I have built something similar for my personal use.

The problem is that some filings especially the 10-K onces, can be very long (too long for chatGPT or Claude). Only Gemini can handle them properly. It can be quite expensive if a user chats a lot with a long filing.

ashr_|1 year ago

Yeah this is true. But the context windows for most of the frontier models have been increasing over time so this will most likely be solved soon. It's good enough to still get a decent amount of value.

flessner|1 year ago

Who's the target audience here? Trading firms and funds have a comparable, likely far better system, already.

Many hobbyists and "algo traders" only care for actionable results. Going through millions of records manually, even with AI, isn't really helpful.

Maybe I am in contact with too many quants, but this isn't even 5% of what's relevant for any given financial decision. Find out what the other 95% are about, integrate them into the product - now you have Bloomberg Terminal, great.

ashr_|1 year ago

Majority of the users are hobbyist value traders and traders at medium-sized firms that spend a lot of time going through 10k's and 10q's manually.