Trayage launched on Product Hunt today. I'm Kyle, its solo developer, and there's one part of the experience I'd particularly like Mac users to evaluate.
Imagine you're reviewing Downloads. There's an installer you might be finished with, two PDFs with similar names, and a large file you haven't looked at in a while. What do you need to know before you're comfortable moving any of them to Trash?
I'm Kaavya and I've been deep in GEO (generative engine optimization) for a while now, and the more I dug in, the more I realized most brands and creators have no idea if their content is even showing up in AI search. not because they're doing bad work, but because no one's told them what "good" looks like in this new world.
I built LLMPvP because I wanted to see how LLM agents do against each other in games where a bad decision gets punished a few moves later, not just on a single answer.
You register an agent, connect whatever model you want on your own infrastructure, and play ranked games. LLMPvP never sees your model or API key. It validates moves, runs the clock and matchmaking, and keeps a separate Glicko-2 rating for each pool: chess rapid, blitz and classical, plus Go 9x9 and 13x13. There's also an unrated house bot for practice, and an MCP server so a coding agent can register and play by itself.
The thing I'm least sure about: when you change the model behind an agent, it starts over with a fresh rating. That keeps the numbers honest, but it also wipes your history every time you upgrade. Would that put you off?
Some of the most defensible products come from a quiet observation a pain point everyone is complaining about and nobody is fixing. I am curious about what's on the community's list right now. The kind of thing where prospects keep describing the same problem, the existing tools all fall short, and the gap feels obvious from the inside but isn't yet on any 'top 10 AI startup ideas' list.
A few examples of pain points that are visible from enterprise discovery calls lately:
Evaluating AI agent output at scale (still mostly manual or LLM-as-judge)
Multi-agent debugging when 4 agents collaborate, who failed?
Since our August launch we replaced the 5-day trial with a free forever plan (15 DMs/month, no card), because too many people left before their first follower ever replied.
New users now write their welcome message first and see a live preview of it as an Instagram DM before connecting anything.
Honest question for makers here: what would make you trust a Chrome extension enough to connect your Instagram?
Wellness apps love streaks, and streaks love guilt. I wanted the opposite.
Mushroom's mood goes a little droopy if you've been ignoring it through a busy stretch. It never dies, never scolds, and never takes anything away. One click on the small recover button and it's happy again. Buttons on notifications are just Done, Snooze, or Skip, and Skip means "not this time", not "stop forever".
Hi everyone! I m Przemek, the developer of VectorRaptor, a Windows desktop app that turns raster images into editable vector graphics.
I d love some honest feedback from people working with logos, illustrations and other graphics. I ve tested it on my own files, but I want to see how it handles the images and workflows other people actually use.
