Not yourmodels,Not your AI.
Decentralised infrastructure for collaborative AI training and inference.
BUILD WITH FLOCK
What is FLock.io?
FLock.io develops AI research and infrastructure for decentralised model training, federated learning, and model inference.
AI Arena
Train and validate AI models through FLock’s decentralised training platform.
Open AI Arena (opens in a new tab)FL Alliance
Refine models collaboratively with federated learning and participants’ local data.
Read the Docs (opens in a new tab)API Platform
Browse supported AI models and connect through one API.
Explore Models (opens in a new tab)FOMO
BetaExplore and trade Real Model Asset tokens for AI models.
Explore Model Tokens (opens in a new tab)Chosen by the most innovative companies
Why Privacy Powers Everything FLock Builds?
Making Decentralised AI Truly Private and Usable
Enable organisations/communities keep their data local, maintain fair value distribution, and launch community-driven AI models that can be deployed and monetised globally.
Built for Privacy, Security, and Scale from Day 1
Secure high-value training with advanced Federated Learning and LoRA-based optimisation, reinforced by a proven on-chain incentive system that ensures resilience, accuracy, and fair rewards across the network.
The Foundation Chosen by World-Class Institutions and Partners
Collaborate with the United Nations Development Programme (UNDP), Alibaba Cloud’s Qwen, and the Ethereum Foundation, providing a privacy-preserving infrastructure that meets real-world security and compliance needs.
Trusted by Industry Pioneers
Read inspiring stories and genuine feedback from users who have experienced exceptional results.
Greg Osuri, Founder of Akash
In just a few clicks, developers can deploy FLock.io Training Nodes and Validators on Akash. We’re thrilled that it’s so easy for developers to participate in decentralized AI development on Akash’s supercloud.
Backed by
and more...
Code Audited by
Stay in touch
News and stories
Why the enterprise AI era is bottom-up, after the top-down SaaS era
Top-down AI pilots, it turned out, were the completely wrong approach. It was the go-to way for organisations to adopt generative AI after it took off. It mimicked the top-down SaaS era, during which the C-suite selected ready-made tools then rolled them out across departments. But they struggled with adoption, with 95% failing to reach production.
FeaturedDecentralized Scaling with FLock.io | FLockCON 2025
Our Chief Scientist Dr.Cheng gave a truly inspiring keynote on decentralized scaling in AI. He unpacked the growing risks of LLM hallucinations and scheming behavior in centralized systems...
FeaturedFLock partners with Qwen - a tier 1 open-source LLM
FLock has partnered with Qwen – a tier 1 open-source LLM, developed by Alibaba Cloud. This is the first time a major Web2 LLM has chosen to deeply integrate with a decentralised training platform...