The context of the AI models is Windows, which indicates the “remembering” information of a model, has increased over time. However there are researchers Proposed New way To enhance the long -term memories of AI models, as they often cannot keep up with several sessions.
Nineteen -year -old founder Dhravya Shah is trying to solve the problems by creating a memory solution in this region, what is called SupermemoryFor AI applications.
Originally from Mumbai, India, Shah started creating consumer-mouth bot and apps a few years ago. He even sold his bot that the formatted tweets were in the handsome screenshots of the social media equipment.
The founder, who was preparing for the entrance test for entry at the IIT (Indian Institute of Technology), made good money from the sale and decided to move to the United States instead to join Arizona State University.
After the transfer, he challenged himself to create something new every week for 40 weeks. In one of these weeks, he created supermemory (which was initially called a context) and kept it Github. At that time, the equipment allows you to chat with your Twitter bookmarks.
The current version of the equipment extracts insights from “memory” or unprotected data and helps the applications better understand the context.

Shah secured an internship in Cloudflair in 2021, where he worked on AI and infrastructure. He later worked as a leading relationship in the organization. At this time, consultants, including CTO CTO Den Nacht, told him to turn the supermmerity into a product.
This year, he decided to create supermemory full -time.
Supermmry, now described as a public memory API for AI applications, creates a knowledge graph based on the data that processes it processed and personalized the context for users. For example, it can help ask for a month -old entry for a writing or journaling app or searching an email app. Since the solution allows for multimodal inputs, it can also allow a video editor to bring relevant resources from a library for a specific prompt.
The company, including files, docs, chat, project, email, PDF and application data stream, says the startup can eat any type of data. Its chatboat and notchakar feature allows users to add memories to the text, add files or links, and to connect to applications such as Google Drive, ODRIVE or concept. There is also a Chrome extension that allows you to easily add notes from a website.
Shah said, “Our main strength is to find insight from any type of structural data and to give applications more about users about it.
The $ 2.6 million dollar seeds led by Supermeri Sus Ventals, Browder Capital and SF1 have secured $ 2.6 million seeds in the round.
Shah said that at one point the Wi -combinator went to him to join him in a batch, but was already investor on his board, so the time was not effective.

“Robot Lawyer” was fascinated by the Acity of the Startup Donotpe, the founder and CEO of Joshua Browder Shah.
“I was attached to Dhravia on the X and what hurt me on the X, how fast he took and made it to me and it persuaded me to invest in his investment,” he said.
The organization has multiple existing customers, with A1 16 Z-Backed Desktop Assistant Kluilly, AI Video Editor Montra, AI Search CategoryOf the composio Multi-MCP equipment rubberAnd the real estate startup RetsThe Also, it is working with a robotics company to hold visual memories captured by a robot.
Although it is acute to consumers, the app makes more understanding about developers’ equipment and it feels more like the playground to use their work or their own apps potentially.
There is enough contestant in the supermemi space in the memory space. Startups (where Shah worked for a while) is creating a memory layer for agents like Feliceis Ventures-Backed Letta and MEM0. Supermemory’s own supporter, Susa Ventures, has invested in memory with Samsung, which can tap in footage for thousands of hours to get insight. Shah says these startups can serve different industries and use cases, the supermemory will stand because it offers less delay.
“More AI agencies will need a memory level. The solution of the supermemi provides you high effectiveness when allowing the surface to surface in the relevant context,” said Browder.
