Launched this week
Open Comet
The autonomous AI browser agent for deep research & tasks
149 followers
The autonomous AI browser agent for deep research & tasks
149 followers
Open Comet is a high-fidelity AI agent that lives in your browser’s sidepanel. Unlike basic assistants, it autonomously browses, researches, and executes multi-step workflows across any website. Built on a "Zero-Data" architecture, it keeps your history local while providing enterprise-grade reasoning. Features include high-fidelity execution guards, STORM-inspired research loops, and full support for both Cloud (BYOK) and Local (Ollama) models.
Open Comet
Hi Product Hunt! 👋 I’m Prince, the creator of Open Comet.
Most AI agents today feel 'trapped' behind a chat box, forcing you to copy-paste data back and forth. I built Open Comet to bridge that gap—an agent that actually lives where you work.
My goal was to combine the power of autonomous research with a strict privacy-first architecture. Whether you're a researcher needing deep link exploration or a developer automating repetitive web tasks, Open Comet is designed to stay out of your data and in your flow.
Excited to hear your feedback and answer any questions!
Autonomous AI research agents are where things are headed for anyone doing serious market analysis. I built PolyMind (https://polyminds.netlify.app/) to track large trades on prediction markets using AI-powered alerts, and the hardest part was always the research layer — synthesizing signals from dozens of sources into actionable insight. An autonomous browser agent that can handle deep research tasks across the web would be a game-changer for financial due diligence workflows. How does Open Comet handle source reliability and conflicting information across pages?
Open Comet
@samir_asadov That’s a really solid use case — PolyMind sounds like exactly the kind of system where this problem becomes critical.
Right now, Open Comet is still early, so I’m not claiming strong reliability yet. It can navigate across pages and extract structured context (DOM + visual cues), but it doesn’t have a mature source ranking or conflict resolution system yet — so in cases of conflicting info, it may still pick the first plausible answer.
That said, this is exactly the layer I’m actively working on:
Adding source reliability scoring (domain trust, repetition across sources, signal weighting)
Building a conflict resolution step where the agent explicitly compares differing claims
Returning evidence-backed outputs (sources + snippets + confidence score)
Introducing a self-verification loop to re-check conclusions
For financial / prediction market workflows like yours, the goal is to move from “automation agent” → research-grade agent that can justify its outputs.
Would love to test this against something like PolyMind once this layer is more stable — feels like a perfect real-world benchmark.
RiteKit Company Logo API
Congrats on the launch! This looks really impressive. I'm curious about the execution guards you mentioned - how do you prevent the agent from getting stuck in loops or taking unintended actions on sites with unusual layouts or dynamic content? Does it have a way to recover gracefully when it encounters something outside its training?
Looks awesome.
On https://opencomet.onrender.com/download it links to clone
git clone https://github.com/princechouhan19/open-comet
but i think the repo has been updated to /opencometAI
Hope this helps!
Open Comet
@jacobah Hey, really appreciate you pointing that out — that’s on me.
You’re right, the correct public repo for now is:
👉 https://github.com/princechouhan19/OpenCometAI
The /open-comet repo is currently private (v3.0 in progress), so that clone link is outdated. I’ll get the website updated.
Thanks again for catching it 🙌
Lovely! Can this browse my logged in pages like linkedin ? And take actions on my behalf. Some UX flows especially linkedin is sub-optimal and this can help save me lot time. Thanks.
Open Comet
@raj_peko Great question, Rajesh 👍
Yes — Open Comet is designed to work on real, logged-in pages (like LinkedIn) and can assist with actions on your behalf.
However, there are a couple of important points:
It can interact with UI elements (clicking, typing, navigating flows) within your active browser session
Since you’re already logged in, it operates within your session context
For sensitive or high-impact actions, it follows a human-in-the-loop approach (asks for confirmation before executing)
So for cases like LinkedIn workflows (posting, navigating profiles, repetitive actions), it can definitely help reduce friction and save time—especially where UX feels sub-optimal.
That said, some flows may vary depending on site restrictions and dynamic UI behavior, but improving reliability across such platforms is an ongoing focus.
Would love to hear your experience once you try it 🙌
Does this work only on Perplexity's Comet browser? Is there an affiliation with them as Open Comet's branding, name and logo is almost 100% on par with Perplexity's - most specifically, their Comet product.
Open Comet
@jacklyn_i Hey Jacklyn, great question—totally fair to ask 👍
No, Open Comet does not work only on Perplexity’s Comet browser, and there’s no affiliation with them.
Tools like Perplexity’s Comet and OpenAI’s Atlas are actually separate AI browsers built from scratch with AI deeply integrated into them. Open Comet takes a different approach—it’s designed to work inside your existing browser (via extension), so there’s no need to switch browsers.
Regarding the naming/branding—yes, it’s inspired by the same emerging category of AI browser agents, but Open Comet is an independent project focused on being more open, flexible, and privacy-first.
Also, one key difference is the approach to data: many AI browsers process and interact with user data at a deeper level to enable automation, whereas Open Comet is designed with a local-first mindset (no data storage/access).
Happy to clarify anything else if you’re curious 🙌
love that you went with local history storage. too many AI tools are black boxes with your data. curious about the STORM research loops - does it actually follow citation trails and cross-reference sources, or is it more like iterative searching?
Open Comet
@piotreksedzik Appreciate that, Piotr 🙌
Great question — the current implementation is closer to iterative research loops rather than a full academic-style citation graph like STORM.
It can:
Perform multi-step searches
Refine queries based on previous results
Aggregate and summarize findings across steps
Cross-referencing does happen to an extent (by comparing results across iterations), but it’s not yet doing deep citation trail tracking or source graph traversal in a strict sense.
That said, moving toward more structured source tracking and citation-aware reasoning is definitely something I’m exploring for future versions.
Would love to hear your thoughts once you try it 👍