I'm the creator of Context Hub. I built it while shipping Rehearsal AI — a voice-powered AI interview-practice platform that's served 5,000+ candidates across IIMs, FAANG prep, and campus placements — and Highlyt, a tool I'm building in stealth.

Most of my work lives at the intersection of AI infrastructure and real product engineering. The demo is always 10% of the work; the other 90% is the boring stuff that makes things actually usable — handling edge cases, building feedback loops, designing systems that degrade gracefully when models hallucinate.

Context Hub came out of frustration. I'd spend twenty minutes making a decision in Claude.ai, then switch to Claude Code in my terminal and have to re-explain the entire context. Multiply by ten decisions a week and you're paying an hour-a-day tax just to keep your AI tools on the same page. So I built the thing that stores the decisions once and lets every AI client read them.

What I'm building

Rehearsal AI

Voice-powered interview practice. Candidates run mock interviews with an adaptive AI that probes deeper when answers go thin — modeling how real interviewers behave. Cross-session memory tracks improvement patterns. Currently serves IIMs, MBA candidates, and tech-role applicants. tryrehearsal.ai

Context Hub

Open-source MCP context layer that Claude, ChatGPT, Cursor, Perplexity, and Claude Code all read from. One npm command stands up a Cloudflare Workers + D1 backed memory server. Free at personal scale. github.com/JaipuriaAI/context-hub

Highlyt

In-progress. highlyt.app

Where to find me

mayankbohra.com — portfolio + writing
github.com/mayankbohradev — open source
@mayankbohradev — short-form thoughts
linkedin.com/in/mayankbohradev — professional updates

Get in touch

Best way to reach me about Context Hub: open an issue on GitHub. For everything else, the channels above.