Teen entrepreneur builds “universal memory” for AI apps with backing from Google, OpenAI, and Cloudflare leaders.

Artificial intelligence models have grown smarter over the years — yet their “memory” remains limited. Despite advances in context windows, AI systems still struggle to retain knowledge across multiple sessions. But 19-year-old Dhravya Shah, the founder of Supermemory, believes he’s found a solution to give AI long-term memory.

From Mumbai to Silicon Valley: The Making of Supermemory

Originally from Mumbai, India, Shah began coding and building consumer-facing bots as a teenager. One of his early creations — a tool that formatted tweets into shareable screenshots — was sold to social media platform Hypefury, marking his first successful exit.

Instead of pursuing his dream of joining IIT (Indian Institute of Technology), Shah used the money to study computer science at Arizona State University in the U.S. Once there, he challenged himself to build a new product every week for 40 weeks. One of those projects would later become Supermemory, first launched as “Any Context” — a simple app that let users chat with their Twitter bookmarks.

Turning an Idea into a Funded AI Startup

While interning at Cloudflare in 2024, Shah worked on AI and infrastructure and later became a developer relations lead. Encouraged by mentors like Cloudflare CTO Dane Knecht, he decided to turn Supermemory into a full-fledged company.

Today, Supermemory is described as a universal memory API for AI apps. It uses a knowledge graph to help applications extract insights from unstructured data, improving personalization and recall. The system can handle multimodal inputs — text, images, and files — and integrate with tools like Google Drive, OneDrive, and Notion.

A built-in chatbot and notetaker let users upload documents, links, and files, while a Chrome extension allows for quick note capture from any website.

“Our core strength is to extract insights from any kind of unstructured data and give apps more context about users,” Shah said. “Since we work across multimodal data, our solution fits AI apps from email clients to video editors.”

$2.6 Million Seed Round Led by Top AI Investors

Supermemory recently raised $2.6 million in seed funding, led by Susa Ventures, Browder Capital, and SF1.vc. The investor list reads like a who’s who of the tech world: Google AI chief Jeff Dean, DeepMind’s Logan Kilpatrick, Cloudflare’s Dane Knecht, Sentry’s David Cramer, and executives from OpenAI, Meta, and Google.

Interestingly, even Y Combinator reached out to Shah, but with investors already on board, the timing didn’t align.

Joshua Browder, founder of DoNotPay and investor through Browder Capital, shared his motivation to back the teen innovator:

“I connected with Dhravya over X, and what struck me was how quickly he moves and builds things. That prompted me to invest.”

Real Customers and Rapid Growth

Supermemory is already being used by a range of companies, including a16z-backed Cluely, AI video editor Montra, AI search tool Scira, and real estate platform Rets. It’s even working with a robotics company to help machines retain visual memories — a key step toward making AI systems more human-like in recall and learning.

Rising Competition in the AI Memory Race

Supermemory faces competition from startups like Letta, Mem0, and Memories.ai, which are also building memory layers for AI agents. But Shah says Supermemory’s low latency and developer-first design make it stand out.

“More and more AI companies will need a memory layer,” said investor Browder. “Supermemory’s solution provides high performance while surfacing relevant context quickly.”

As AI continues to evolve, Supermemory might just become the missing link — giving artificial intelligence a way to truly remember.