The reference window of the AI model, which indicates information for a model’s “memory” information, has increased over time. However, researchers have suggested New ways To increase the long -term memory of the AI model, because they cannot often reference to several sessions.
Nineteen -year -old founder Dhravya Shah is trying to solve problems in the region by constructing a memory solution, which is called SupermarryFor AI apps.
Originally from Mumbai in India, Shah started building consumer-support bots and apps a few years ago. He even sold his bot, which gave the appearance of a tweet in a good looking screenshot in the social media tool hypfari.
The founder, who was preparing for an entrance exam to enter the IIT (Indian Institute of Technology), earned good money from this sale and instead decided to go to America to attend Arizona State University.
After transferring, he challenged himself to create something new each week for 40 weeks. During one of those weeks, he created Supermemiry (which was initially called in any context) and placed it on git. At that time, the tool allowed you to chat with your Twitter Bookmark.
The current version of the tool gives insight by “memories” or unnecessary data and helps applications to understand reference better.
Shah acquired an internship in Claudflair in 2024, where he worked on AI and infrastructure. He later worked as a developer Relations lead in the company. During this time, advisors, including Cloudflare CTO Dane Knecht, asked him to replace supermmori in a product.
This year, he decided to make Supermmori full -time.
Supermarm, which is now described as a universal memory API for AI apps, makes a knowledge graph based on data processes, and individualizes references to users. For example, it can support the query in a month old entries for a writing or jernling app, or discover the email app. As the solution allows for multimodal input, it may also allow a video editor to bring relevant property from the library for a particular prompt.
Startup can ignite any type of data, the company says, including files, doors, chats, projects, email, PDF and app data stream. Its chatbott and notakar feature users allow users to add memories, add files or links and connect to apps such as Google Drive, Onedrive, or Dharna. There is also a chrome extension that lets you easily add notes to a website.
“Our main strength is to extract insight from any kind of unnecessary data and give more reference to apps about users. As we work in multimodal data, our solution is suitable for all types of AI apps from email clients to video editors,” Shah said.
Supermemer has achieved seed funding of $ 2.6 million led by Susa Ventures, Bronner Capital and SF1.VC. This era also includes individual investors such as Claudflare’s Canchet, Google AI chief Jeff Dean, Deepmind Product Manager Logan Kilpatrick, Sentry founder David Cramemer and Openai, Meta and Google officials.
Shah said that at one point the Y Combinator also approached him to join one of his batch, but he already had investors, so did not work on time.
“Robot lawyer” was influenced by Shah’s tenacity, the founder and CEO of Startup Donotpe, who runs the brilliant Capital as a single GP.
He said, “I am associated with Dhravya on X, and what I was killed was how quickly she walks and manufactures things, and inspired me to invest in it,” she said.
The company has many current customers, including A16Z-supported desktop assistant CLUELY, AI Video Editor Montra, AI search MalevolentComposes Multi-MCP Equipment RubAnd real estate startup RetainesIn addition, it is working with a robotics company to maintain visual memories captured by a robot.
Although it is a slant towards consumers, the app feels more like a playground for developers to understand more about the toilet and potentially uses it in its workflow or your own app.
There are enough contestants in memory space in Supermeri. Fellicis Ventures-supported Lata and MEM0 (where Shah worked for a while) are constructing a memory layer for startup agents. Susa Ventures has invested in memories with Samsung, his backcover, Susa Ventures of Supermory. Shah says that while these startups can serve various industries and use matters, supermimeries will stand out as it provides less delay.
“More and more AI companies will require a memory layer. The solution of supermarmori provides high performance, while allowing you to take the relevant reference to the surface quickly,” the brown said.
