Skip to main content

How it works

The toda MVP has four parts.

1. Capture​

Bring in the AI data you already create.

For consumers, the browser extension can capture conversations from supported AI products. Businesses and developers can connect through the API or SDK. Agents can write memory directly.

The source changes. The network does not.

2. Save​

toda turns captured data into persistent memory tied to its owner.

The raw content stays offchain and private. Onchain records are used where public coordination matters, such as ownership, staking, and access rules.

You are not locked into the AI product that created the data.

3. Validate with staking​

People can stake behind data they believe is useful.

Staking creates an economic signal. More support can make valuable data easier to identify and harder to spam. It does not certify truth, replace review, or make every popular claim correct.

We use S as the shorthand for this signal.

4. Use useful data onchain​

Data with a strong staking signal can enter the network's useful layer. Ownership, permissions, and exchange can be managed onchain while the private content stays protected.

This is the foundation for shared datasets, portable personal memory, business knowledge, and memory that software can use across products.

The same layer can later power deeper model integrations and a memory exchange where people share, license, or sell useful data on their own terms.

your AI data
     |
  capture
     |
 private memory
     |
 staking signal (S)
     |
 useful data managed onchain

The boundary that matters​

toda is not putting everyone's private conversations on a blockchain.

The private data and the public coordination layer have different jobs. Content stays protected. The network handles the signals and rules that let people use that content together.

Next: For consumers or Why toda?.