Hacker News new | ask | show | jobs
Show HN: Parallel AI agents that research a stock simultaneously (dapto.ai)
1 points by sharmasachin98 116 days ago
Hi HN,

I’ve been working on a system that runs multiple AI agents in parallel to perform structured research instead of generating a single summary response.

One use case I tested recently was stock research.

When you properly research a stock like NVIDIA, you usually open multiple tabs:

- Financials - Earnings reports - Analyst sentiment - Competitors - Recent news - Risks - Market positioning

Most AI tools generate one combined answer, which often becomes shallow or blended.

So I built a workflow execution agents that:

- Spawns multiple specialized agents at once - Assigns each agent a focused responsibility (financials, competitors, risks, etc.) - Runs them in parallel - Normalizes structure - Compiles everything into a single structured research report

Instead of one AI response, you get multiple independent research threads that are merged into a coherent output.

The goal isn’t “better summaries.” It’s structured multi-angle research without manually orchestrating prompts.

Here’s a short demo using NVIDIA stock:

https://youtu.be/QBmFK843Kuo

Would love feedback on:

- Does parallel specialization meaningfully improve depth vs single-thread LLM prompts? - Where else would this model be more useful (beyond stock research)? - What would you want to see measured (quality benchmarks, latency, cost breakdown)?

Happy to answer technical questions.