← Back to Home

General

What are MCP and x402?
MCP (Model Context Protocol) allows AI models to securely access local tools. x402 is a standard for web micropayments. Payflow uses both to enable agentic commerce.
What is cryo?
Cryo is an open-source tool for extracting blockchain data at scale. It's designed to be fast and efficient, allowing you to get large datasets like blocks, transactions, and logs quickly. Payflow uses an MCP server around Cryo to provide data to AI models.

Payflow

What blockchains are supported?
Currently, the Cryo MCP server supports Ethereum Mainnet. We are working on adding support for other blockchains. Payments are settled on Base in USDC. If you have a specific blockchain you would like to see, please let us know.
What rate limits are in place?
The Cryo MCP server limits the range per query to 10,000 transactions or 10,000 blocks for all datasets, with a fixed cost of 0.05 USDC per request.
Why I am required to pay for the data? How is pricing determined?
The cost of the data is determined by the size of the query. The more data you request, the higher the cost. The cost is paid in USDC, and is relative to the number of transactions or blocks in the dataset. Currently, the cost is 0.05 USDC per 10,000 transactions or 10,000 blocks.

Note: The cost is currently fixed, and is not based on the complexity of the query or the dataset, or on demand. This is because we use the `exact` scheme of the x402 protocol, which is the only scheme that is implemented in the x402 library so far.
What cryo datasets are available?
The Cryo MCP server supports all datasets available in the Cryo library.
Is it safe to put my private key in the configuration?
The Cryo MCP server runs entirely on your local machine. Your private key is stored locally and never transmitted over the network. It is only used to sign payment transactions on your behalf, as requested by the AI model you are interacting with. You can also configure the Payflow MCP server with hard limits per query. Nevertheless, ALWAYS use a new burner wallet when setting it up!
Where does the data come from and how is it stored?
The data is sourced from an archive reth node run by Chainbound. Once the dataset is generated, they are stored in Parquet files on the remote server, and retained for 12 hours.