Taranjeet Singh (pictured above, right) has started six companies, some of which have failed and others with varying degrees of success. His seventh Mem0 could be his defining one.
The startup starts with the premise that large-scale language models cannot remember past interactions like humans. If two people are chatting and the connection drops, you can resume the conversation. In contrast, an AI model forgets everything and starts from scratch.
Mem0 fixes that. Singh calls this a “memory passport,” where the AI’s memory moves across apps and agents, much like email and logins do today. The YC-backed startup launched in January 2024 and has raised $24 million ($3.9 million in undisclosed seed funding and $20 million in Series A).
Basis Set Ventures, an early-stage AI-focused fund, led the Series A, with participation from existing investors Kindred Ventures and Y Combinator, as well as new backers including Peak XV Partners and GitHub Fund.
Notable angels include Dharmesh Shah (HubSpot), Scott Belsky (ex-CPO Adobe), Olivier Pomel (Datadog), Thomas Dohmke (ex-CEO GitHub), Paul Copplestone (Supabase), James Hawkins (PostHog), Lukas Biewald (Weights & Biases), Brian Balfour (Reforge), and Philip Rathle. (Neo4j), and Jennifer Taylor (former president of Plaid).
The bets on Mem0 (pronounced “mem-zero”) by several of the leaders who helped shape the modern software ecosystem highlight its promise, and the traction of the four-person team backs that up.
The open source API, which claims to be the most widely adopted memory framework by AI developers, has over 41,000 GitHub stars and 13 million Python package downloads to date. In Q1 2025, Mem0 processed 35 million API calls. By the third quarter, that number had jumped to 186 million, an increase of about 30% month-over-month.
tech crunch event
san francisco
|
October 27-29, 2025
Beyond open source adoption, over 80,000 developers have signed up for its cloud services. Mem0’s cloud API can now handle more memory operations than other providers and serves as the dedicated memory provider for AWS’ new agent SDK.
In early 2023, Singh was still in Bangalore, India. He started his career as a software engineer at Paytm, one of India’s most valuable startups, before becoming Khatabook’s first Growth Engineer. He left in late 2022, just as the ChatGPT wave was reaching its peak, and built one of the first GPT app stores to reach over 1 million users.
That experience inspired him to create Embedchain, an open source project that allows developers to index, retrieve, and sync unstructured data. Once the project took off and garnered over 8,000 GitHub stars, Singh sent over 200 cold emails to Silicon Valley founders, investors, and engineers.
“I reached out to almost every famous technology entrepreneur you’ve ever heard of, and they were very persistent. Several of them responded and, after hearing our story, scheduled us to fly from Bangalore to San Francisco within 36 hours,” Singh said.
Upon arriving in the US, Singh reunited with his long-time friend and now co-founder and CTO, Deshraj Yadav, who led the AI platform at Tesla Autopilot. They previously built EvalAI together, an open source Kaggle alternative that grew to 1.6K GitHub stars.
While experimenting with Embedchain, the duo launched a meditation app inspired by Indian yogi Sadhguru. As the app quickly spread in India, Singh said users continued to share the same feedback. “Hey, I’m on a meditation journey and the app doesn’t remember it.” So they pivoted from Embedchain to Mem0 to solve that problem.
The idea of memory in AI is not new, but it is quickly becoming an important battleground. For example, OpenAI began testing ChatGPT’s long-term memory capabilities in early 2024, and its CEO Sam Altman has indicated that persistent memory will be central to OpenAI’s next hardware device. Other AI labs are also launching experimental memory systems for agents.
Singh argues that while major AI labs are building memory systems, there is little incentive to make them portable or interoperable. “Memory is becoming one of the key moats now that LLM is becoming a commodity,” he said.
He explains that while consumers can enjoy a persistent, personalized experience with ChatGPT, developers who want to build applications (for example, a financial companion that remembers a user’s transaction history) need an open, neutral solution like Mem0.
“We want to enable developers to deliver day-one personalization through a shared memory network,” said Singh. “Think of this as Plaid for memory. This is the second act. For now, we’re focused on building the best memory products possible.”
Mem0’s framework allows developers to store, retrieve, and evolve user memory across models, applications, and platforms. It is model agnostic, compatible with OpenAI, Anthropic, or other open source LLMs, and integrates directly with frameworks like LangChain and LlamaIndex.
Developers use Mem0 to create applications that get smarter with each interaction: therapy bots that remember past conversations, productivity agents that remember personal habits, and AI companions that adapt over time. Our customers range from indie developers to enterprise teams building copilot and automation tools.
“Memory is the foundation of the future of AI, which is why we backed Mem0 in its early days, even before YC,” said Lan Xuezhao, founder and partner at Basis Set Ventures. “We are excited as our team continues to tackle one of the most difficult and important infrastructure challenges: enabling AI systems to build persistent, contextual memory.”
Other early-stage startups in the memory space include Supermemory (whose founder briefly worked at Mem0), Felicis-backed Letta, and Memories.ai.
