Hi Product Hunt, I’m Sulabh. AI kept forgetting my projects, so I built a memory for it.

Hi everyone, I’m Sulabh, based in India.

I’ve spent much of my career working around large transformation programs: enterprise systems, manufacturing, supply chains, data, and AI governance. Alongside that work, I enjoy building practical AI products and learning in public.

Rta-Smriti Brain began with a very ordinary frustration. Every new coding-agent session meant explaining the same project again: the architecture, decisions, failed approaches, release constraints, and what had actually been verified.

So I built a local, open-source project memory for Codex, Claude Code, Cursor, and other coding agents. It maps a repository, connects decisions with evidence, checks whether the context is still fresh, and prepares a focused context pack for the next task. It runs locally, with no hosted account or telemetry.

It’s currently at v0.4.0-alpha, and I’m looking for honest tests more than polite applause.

If you work with AI coding agents, what project context do they forget most often?

Rta-Smriti Brain:

https://www.producthunt.com/products/rta-smriti-brain?launch=rta-smriti-brain