OnDB

What kind of data can I publish?

Any REST or GraphQL API, PostgreSQL, MongoDB, Supabase, ClickHouse, or hosted dataset. If it has a schema, it works — the OnDB connector wraps it into a queryable endpoint.

How do payments work?

OnDB implements HTTP 402 natively. Every request is priced per query and settled automatically via x402 or Machine Payments Protocol (MPP) — no billing logic or reconciliation on your end.

How does revenue sharing work?

You set your per-query price in code. OnDB meters usage, splits revenue between providers and the platform automatically, and pays out to your wallet. Share ratios are configured server-side.

How do AI agents discover my data?

Agents read a standardized skills.md definition exposed by your OnDB endpoint. No marketplace listing — just publish, and any agent that speaks the protocol can find, query, and pay for your data.

How is this different from other AI tools?

Most tools help agents generate output. OnDB is the infrastructure between agents and the data they need — a payment-native query layer that routes across providers and settles in real time.

Does my data leave my infrastructure?

No. The OnDB connector runs inside your environment and queries are processed locally — only the query response is returned to the agent. Your databases, APIs, and raw data stay exactly where they are.

How is data secured in transit?

All traffic is TLS 1.3 end-to-end. Payment proofs are signed via x402, and data never leaves your infrastructure — OnDB routes requests to your connector without proxying payloads through a middleman.