Taranjeet Singh (pictured above, right) has launched six companies, with some failing and others seeing varying degrees of success. His seventh, Mem0, could be his defining one.
The startup starts with the premise that large language models can’t remember past interactions the way humans do. If two people are chatting and the connection drops, they can resume the conversation. AI models, by contrast, forget everything and start from scratch.
Mem0 fixes that. Singh calls it a “memory passport,” where your AI memory travels with you across apps and agents, just like email or logins do today. The YC-backed startup, launched in January 2024, has raised $24 million ($3.9 million in previously unannounced seed funding and a $20 million Series A).
AI-focused early-stage fund Basis Set Ventures led the Series A, with participation from existing investors Kindred Ventures, which led its seed round, and Y Combinator, as well as new backers including Peak XV Partners and the 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), Philip Rathle (Neo4j), and Jennifer Taylor (former president, Plaid).
Having several leaders who helped shape the modern software ecosystem bet on Mem0 (pronounced “mem zero”) underscores its promise, and the traction from the four-person team backs it up.
So far, the open source API, which claims to be the most widely adopted memory framework for AI developers, has surpassed 41,000 GitHub stars and recorded over 13 million Python package downloads. In Q1 2025, Mem0 processed 35 million API calls. By Q3, that number jumped to 186 million, growing roughly 30% month over month.
Techcrunch event
San Francisco
|
October 27-29, 2025
Beyond open source adoption, more than 80,000 developers have signed up for its cloud service. Mem0’s cloud API now handles more memory operations than any other provider and serves as the exclusive memory provider for AWS’s 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 quit in late 2022, just as the ChatGPT wave was about to crest, and built one of the first GPT app stores, which scaled to over a million users.
That experience led him to create Embedchain, an open source project that lets developers index, retrieve, and sync unstructured data. As the project took off, earning more than 8,000 GitHub stars, Singh sent over 200 cold emails to founders, investors, and engineers in Silicon Valley.
“I reached out to almost every famous tech entrepreneur that you might have heard of and was quite persistent. Some of them responded, and after hearing us out, scheduled us to fly from Bangalore to San Francisco within 36 hours,” Singh said.
Once in the U.S., Singh reconnected with his longtime friend and now co-founder and CTO, Deshraj Yadav, who had led the AI Platform at Tesla Autopilot. Together, they had previously built EvalAI, 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. The app went viral in India, but Singh says users kept sharing the same feedback: “Hey, I’m on this meditative journey, but the app doesn’t remember that.” So they pivoted from Embedchain to Mem0 to solve that problem.
The idea of memory for AI isn’t new, but it’s quickly becoming a critical battleground. OpenAI, for instance, began testing long-term memory features in ChatGPT in early 2024, and its CEO, Sam Altman, has hinted that persistent memory will be central to OpenAI’s upcoming hardware device. Other AI labs are also launching experimental memory systems for their agents.
Singh argues that while big AI labs are building memory systems, they have little incentive to make them portable or interoperable. “Memory is becoming one of their key moats now that LLMs are getting commoditized,” he said.
He explains that while consumers can enjoy persistent, personalized experiences in ChatGPT, developers who want to build applications — say, a finance companion that remembers a user’s trading history — need an open, neutral solution like Mem0.
“We want developers to offer day-one personalization through a shared memory network,” Singh said. “Think of it as Plaid for memory. That’s act two. For now, we’re laser-focused on building the best memory product possible.”
Mem0’s framework lets developers store, retrieve, and evolve user memory across models, applications, and platforms. It’s model-agnostic, compatible with OpenAI, Anthropic, or any open source LLM, and integrates directly with frameworks like LangChain and LlamaIndex.
Developers use Mem0 to create applications that grow smarter with every interaction: therapy bots that recall past conversations, productivity agents that remember personal habits, and AI companions that adapt over time. Customers range from indie developers to enterprise teams building copilots and automation tools.
“We backed Mem0 from its earliest days — even before YC — because memory is foundational to the future of AI,” said Lan Xuezhao, founder and partner at Basis Set Ventures. “We’re doubling down as the team continues to tackle one of the hardest and most important infrastructure challenges: enabling AI systems to build lasting, contextual memory.”
Other early-stage startups in the memory space include Supermemory (whose founder briefly worked at Mem0), Felicis-backed Letta, and Memories.ai.
