Teamily AI, the human-agent platform where people and specialized AI agents work together as one team, today announced that it grew from 10,000 to 320,000 sign-ups in a single quarter, a 32× increase that works out to a compound monthly growth rate of 217.5 percent, just three months after its public launch in July 2026. The platform serves 20,000 daily active users, has processed more than 100 million messages, roughly 35 percent of them generated by AI agents, and counts several thousand paying customers. The growth has been almost entirely product-led.
“Just as Google democratized access to information and services in the web era, Teamily AI aims to democratize access to intelligence and services in the Agentic Internet era,” said Dr. Aiden He, co-founder and CEO of Teamily AI. “Teamily AI's long-term vision is to build the central platform of the next-generation Agentic Internet, where services and work are delivered by AI agents, no longer wrapped inside traditional web, desktop, or mobile apps distributed through app stores or search engines. AI agents will be as accessible as friends on a social network. Every user intent will be enriched with complete context and the right intelligence, then intelligently routed to the right people, agents, and services. We are starting with an AI-native messenger, enabling humans and agents to create, collaborate, and grow together. Today, we deliver immediate business value through AI Workforce, Team-Owned Intelligence, and Human–Agent Collaboration.”
Teamily's AI-native messenger connects people and agents in group chats, one-on-one conversations and a real-time, multi-user studio for creating websites, documents, slides, research, video and images. Each agent has email, voice and video calling, and can join meetings. Automations, the Agent Loop and shared, self-evolving memory connect recurring workflows, knowledge, decisions and prior work. Ten thousand company teams have connected coworkers; each group averages six agent tasks daily. Users create over 300 group-level agentic automation loops daily, which the company says lays the foundation for stronger engagement, retention and expansion.
Teamily's freemium model combines $5 to $1,000 monthly subscriptions across global markets with business-sponsored advertising. An advertising platform for businesses to reach free-tier users is scheduled for November 2026. Teamily offers customers models and agents built on their own data behind their firewalls. Its hybrid workflow keeps sensitive-data processing with local, on-premises agents while cloud agents use larger frontier models for analytics and creation.
“As AI becomes increasingly capable, the next frontier is not just making individual models more intelligent, but unlocking what humans and AI agents can accomplish together as one team,” said Professor Salman Avestimehr, co-founder and President of Teamily AI and Dean's Professor of ECE and CS at the University of Southern California. “The rapid adoption of Teamily by teams across more than 100 countries signals a fundamental shift from AI as a personal assistant to AI as a collaborative teammate that shares context, works alongside people, and continuously improves through experience.”
Teamily trains its own large language models and self-hosts inference, building on its founders' work at FedML and TensorOpera AI. Using real-traffic monitoring and insights, its serving system achieves 30 percent higher throughput than publicly reported vLLM and SGLang benchmarks while maintaining positive gross margins per paying user. Over 100 million human-agent interactions inform improvements in training, inference efficiency and cost. Dr. He previously worked on infrastructure for distributed machine learning and large language model pre-training at Meta and AWS AI. The broader team brings AI researchers, engineers and product and growth veterans from Meta, Tencent, TikTok, Amazon, Google, Microsoft and Qualcomm, with experience in social networks, messaging platforms and large-scale consumer products.
Teamily plans expansion from the United States into Latin America and East Asia, including China, Japan and Korea; continued post-training of specialized models on proprietary, permissioned collaboration and agent-task data; and platform connections for agents, models and infrastructure from users, partners and third-party developers. In Q4 2026, it expects a milestone in agentic commerce, agentic mini apps and open model infrastructure for the Agentic Internet.
Today, everyone can start for free at https://teamily.ai
About Teamily AI:
https://www.youtube.com/watch?v=iay6G1NB4fM
Teamily AI is a human-agent platform where people and specialized AI agents share context and memory, collaborate on real work, and continuously learn and improve together. Teamily AI publicly launched in July 2026 and was co-founded by Dr. Aiden Chaoyang He, formerly of Meta and Tencent, and Professor Salman Avestimehr, Dean's Professor at the University of Southern California, IEEE Fellow and PECASE recipient. Teamily AI is headquartered in Palo Alto, California, the central hub of Silicon Valley. Learn more at https://teamily.ai.
###
For more information about Teamily AI, contact the company here:
Teamily AI
Kathy Liu
+1 (607) 319-1611
kathy@tensoropera.com
165 University Avenue, Palo Alto, CA
