HN Explorer

Show HN: Autonomous AI agents that monitor the stock market for you (decodeinvesting.com)

40 points by clark-kent ↗ HN
We created autonomous AI Agents that monitor the stock market for you while you go about your day.

How it works: Tell our AI Assistant what you want to monitor, and it creates a project for our team of autonomous AI Agents. You'll get notifications (email + app) when significant events matching your criteria are detected. For short-term projects, you'll be notified when your analysis is ready.

Behind the scenes: When you give the AI Assistant a request to monitor an entity (like a stock or group of stocks), an AI Project Manager plans the project and breaks the project down into manageable tasks. These tasks run asynchronously - some recurring (hourly/daily/weekly/monthly/quarterly/yearly), others one-time.

Example prompts you can try: Long-term monitoring: - "Monitor Apple stock and notify me of any important events and red flags" - "Monitor Apple, Google, Microsoft, and Meta stock. Notify me if any of them start trending toward being undervalued"

Short-term analysis: - "Create a project to analyze the last 30 earnings calls for Tesla, spot trends, and how the business has evolved over time"

You can track the progress of all tasks as the AI Agents work in the background.

Try it here: https://decodeinvesting.com/chat

This is still an early version - we're actively improving it based on feedback. Would love to hear what you think and what features you'd want to see next!

Previously shared our AI-powered Stock Market Research Analyst: https://news.ycombinator.com/item?id=41156478

32 comments

[ 4.9 ms ] story [ 86.1 ms ] thread
htrp
Is chat really the best interface for a stock market analysis tool?
great_psy
When you say, “monitor red flags” what is actually happening ?

How are you determining red flags ?

Will an LLM be running every x hours analyzing news, so I can give very specific things to look for ?

clark-kent
Red flags are things like significant drops in stock price, revenue, earnings, customer satisfaction, lawsuits, or changes in senior management.

You can specify the red flags, or the LLM will specify them for you don't include them.

You can also specify how often you want the LLM to monitor things for you.

citadel_melon
What data do you collect about users, and are there any protections against taking trading ideas? Institutional traders would be wary of using the application, cautious of losing any edge by tipping their hands of what equities they are looking at to a third party.

If the application is meant for retail, how do you plan on funding the project? Answering this question would better illuminate the stakeholder's incentives and if they align with users' interests.

clark-kent
We have offerings for both retail and institutions. Institutions can choose between on-prem solutions or a dedicated cloud deployment.

Both retail and institutions can also use for our SAAS solution. We don't share our user data with third parties.

I do review the AI interactions for the following reasons:

- To ensure the AI system (AI assistants and agents) meets expectations.

- Safety: To ensure the AI system is not generating false or misleading information.

- To understand how the AI system is performing and how to improve it.

rybosworld
Locking this behind a signup form seems like a mistake.
TripleChecker
I tend to agree. If you're dead set on locking behind a signup, at least the example queries shown (e.g. 'Summarize Apple's latest earnings' should be clickable so that users can see how it works.
rvz
This looks like a scam.
bhhaskin
If it looks like a duck and walks like a duck....
clark-kent
It's just a research and analysis tool that automates the process of reading lots of SEC Filings and earnings call transcripts and checking the news and stock quotes back to back.

Just automates the research that users would do manually.

thatguysaguy
> Monitor Apple, Google, Microsoft, and Meta stock. Notify me if any of them start trending toward being undervalued

If you actually had a system that could detect when a stock was "undervalued" you should just trade directly on that system instead of selling it to consumers. This is pretty clearly scammy/predatory towards retail investors.

clark-kent
We only look at when the stock is objectively "undervalued". I'm not saying the system will tell you that a stock like Palantir is undervalued because it has a bright future. No. This system only tells when a stock is objectively undervalued. There are a few ways to know this.

Here some examples:

- Market cap is less than book value.

- Market cap is less than free cash flow x 8

There are other ways, but these are some examples. Being undervalued doesn't mean its a good company, the trick is knowing why its undervalued and if that valuation is justified.

m3kw9
If any of these metrics were arbitrage, it would mean the fastest execution wins and the gap wouldnt last long. These metrics almost never work
qeternity
> the trick is knowing why its undervalued and if that valuation is justified.

If it’s justified, then it’s not undervalued.

Everything you described above is also just wrong and you should please stop encouraging uneducated retail investors to believe anything that fits these heuristics is “objectively undervalued”.

danielmarkbruce
> This system only tells when a stock is objectively undervalued. There are a few ways to know this.

> - Market cap is less than free cash flow x 8

Lol. "Objectively".

mnky9800n
If that were true you wouldn’t need an ai to tell you.
stormfather
Not necessarily. Say you can make 20% a year with your system. That's very respectable. But unless you have a large account already (1M?), you will make more by providing this knowledge to others.
woleium
You don’t need 1M to trade 1M. Ive seen retail accounts with 50x leverage
ankeshk
Can it help with technical analysis as well? And screening of stocks based on technicals?
dboreham
You'd need to figure out an embedding for the seaweed.
clark-kent
Not yet. I plan to see if we can offer value for technical analysis. We specialize in fundamental analysis.
shadowerm
I haven't really used agentic AI yet and I am already completely sick of it.

40 questions to Saturday just sounds ridiculous.

I feel like on this trajectory my broker is soon going to be pretending my trailing stop is an AI agent monitoring when to get out of the trade "in the background".

itishappy
Is this investment advice?

> "Investment adviser" means any person who, for compensation, engages in the business of advising others, either directly or through publications or writings, as to the value of securities or as to the advisability of investing in, purchasing, or selling securities, or who, for compensation and as part of a regular business, issues or promulgates analyses or reports concerning securities...

https://www.govinfo.gov/content/pkg/COMPS-1878/pdf/COMPS-187...

whycome
Does it matter here?

> any person

The AI defense.

cyanydeez
The person who made the tool is culpable. Same way if a lawyer hires an intern that preps legal documents is responsible.

However, governments in thw US love to whitewash liability so this is just academic.

itishappy
Of course it matters! AI doesn't automatically absolve you from your legal responsibilities...

I'd bet there's a person somewhere that receives money in exchange for sharing the AI's analysis, which would fall squarely into the second half of the definition:

> [any person] who, for compensation and as part of a regular business, issues or promulgates analyses or reports concerning securities...

nerdjon
Is there any custom models or logic to this; or is this just a cron job wrapped around an LLM?

Capturing data, feeding it to an LLM, and hoping it actually does comparisons properly?

Is the LLM writing code that is being executed for each query or job?

Otherwise, all I am seeing here (and based on your comments saying "LLM") it just appears this grabs data, gives it to an LLM, and "magic output"

clark-kent
There is a lot of custom logic and proprietary data, including a custom LLM framework and an agentic workflow framework as well as an insight engine.
YaBa
Eh... doesn't have data after October 2023.
clark-kent
The cutoff date for the data used to train the model is October 2023, but the AI Assistant can access historical, real-time market data, search, and browse the web. It's still not perfect, though.