Coro is a global conversation space where you speak naturally and relevant voices find you, without rooms, topics, follower graphs, or endless searching. Conversations cross languages: every voice is rendered in a language you understand, while replies bring the context you need. Ambient, explainable routing helps people find one another instead of creating another feed.

How did GPT-5.6 change the ambition or scope of what you shipped?

I sketched an early idea with GPT-5.5 and Opus but didn't like the results, then I revisited the project with GPT-5.6 and it helped me go through a lot of uncertainty and it even helped me devise several simulation and experimentation harnesses that I could run locally in a loop that shaped the final version that I'm shipping. I was also impressed by the parallelization, I could spawn multiple different coding sessions in Codex and they would coordinate automatically, even with many multiple subagents. I have written 0 lines of codes for this.

Hey Product Hunt 👋 I built Coro because online communities increasingly feel like places to publish at people, rather than places to meet them. The usual choices are noisy global chats, isolated topic rooms, or algorithmic feeds optimized for passive consumption. Coro starts with a simpler idea: say what you're thinking, and your voice should find someone. There are no rooms, hashtags, or follow graph. Relevant voices arrive in your personal Stream, conversations cross languages, and missing context comes along so you can reply naturally. The question behind this launch is whether a global public conversation space can create meaningful two-way interactions, especially between people who might never otherwise find one another. I’d love your honest feedback: what would make a space like this feel human, relevant, and safe enough to return to?

How did GPT-5.6 change the ambition or scope of what you shipped?

Coro

Your voice finds someone

Your voice finds someone