Vana

About Vana

What Is Vana?

Vana is a decentralized network designed to give users full ownership and control of their data. Originating from MIT research, the project enables individuals to contribute their data non-custodially to AI systems, forming a foundation for open, community-driven AI development. Vana is open-source, permissionless, and aims to replace centralized data models with decentralized alternatives.

Mission and Vision

Vana envisions a digital future where users—not platforms—own the datasets that power artificial intelligence. The goal is to create a user-owned AI foundation model by enabling mass data contribution from individuals and communities, removing data from walled gardens, and placing it back in the hands of those who generate it.

How Vana Works

Non-Custodial Data Infrastructure

Vana allows users to export and contribute data without giving up control. This non-custodial approach ensures privacy, transparency, and sovereignty over how data is used and where it’s stored. Every user becomes a stakeholder in the AI systems they help build.

Open and Decentralized Protocol

Operating as a decentralized protocol, Vana supports the formation of Data DAOs—community-run data collectives that manage, curate, and monetize datasets. This model shifts data governance from corporations to individuals and independent groups.

Building with Vana

Starting a Data DAO

Through Vana’s infrastructure, anyone can launch a Data DAO. These decentralized organizations collect and manage specialized datasets (e.g., resume data, social media content) with transparent governance and shared incentives. Vana provides documentation, tooling, and support for these initiatives.

Running a Node

Developers and technologists can run Vana nodes to help support the network. Nodes validate transactions, host data, and facilitate the decentralized architecture that makes user-owned AI possible.

Applications and Use Cases

Training Foundation Models

Vana enables the creation of user-owned foundation models. With access to diverse, high-quality datasets sourced directly from users, these models reflect real-world input rather than limited proprietary corpora. The ultimate aim is to build a global-scale foundation model shaped by 100 million contributors.

Personal AI Systems

Vana also supports local deployment of LLMs (large language models) trained on individual data. This allows users to run their own personalized AI agents, capable of understanding their preferences and context without compromising privacy.

Vana Ecosystem

Data Hub and Cohorts

The Vana ecosystem includes multiple active cohorts such as the Aurora and Blink initiatives, which organize early adopters, developers, and researchers. These cohorts contribute to Data DAOs, experiment with personal server setups, and share insights on user-owned AI practices.

Community and Governance

Governance on Vana is decentralized. Data DAOs and contributors play an active role in shaping network rules and protocols. Through the Open Data Foundation and partner organizations like Corsali, community members help expand the mission of open infrastructure and user data rights.

Milestones and Innovation

Technical Achievements

• 2021: First onchain training dataset
• 2022: Patent for non-custodial data systems
• 2023: Launch of local LLMs trained with personal data
• 2024: Deployment of first Data DAOs and Satori testnet

Long-Term Goals

Vana is scaling toward onboarding 100 million users, establishing global data liquidity, and aggregating the largest user-contributed training dataset in history. This will support the training of user-owned foundation models that rival the capabilities of centralized AI systems.