Morpho Agents for lending data
DIA Oracles showcased Morpho Agents, an infrastructure layer that lets AI agents query lending markets and vaults directly to obtain machine‑readable market data. The capability is positioned to provide programmatic access for automated agents that need real‑time lending signals. (x.com) (x.com)
Crypto lending runs on software pools, not bank balance sheets: depositors supply tokens, borrowers post collateral, and smart contracts set the terms. Morpho and DIA are now pitching tools that let artificial intelligence agents read those markets directly instead of scraping dashboards built for humans. (morpho.org) (diadata.org) Morpho said on April 8 that its beta “Morpho Agents” stack gives agents read, simulate, and write access to Morpho protocols on Ethereum and Base through a command-line interface, a Model Context Protocol server, and a protocol knowledge base. Morpho said the product starts with a User Agent for transactions and a Builder Agent for coding integrations. (morpho.org) Morpho’s own documentation says its GraphQL application programming interface already exposes live and historical data on markets, vaults, annual percentage yields, total value locked, utilization, fees, allocations, rewards, and user positions. The same docs say the service is rate-limited at 5,000 requests per five minutes and is provided without a service-level agreement. (docs.morpho.org) A lending market is the basic pool where one asset is supplied and another is posted as collateral. A vault is a managed basket that moves deposits across markets, and Morpho says Vault V2 can route funds through adapters that report asset values in real time. (docs.morpho.org 1) (docs.morpho.org 2) DIA is selling the missing piece for assets that do not trade cleanly on exchanges. Its lending and vaults page says yield-bearing tokens, structured vault tokens, and illiquid positions often need “fundamental valuation” computed from onchain data rather than a last-traded market price. (diadata.org) That matters because an automated agent deciding where to lend or what collateral to accept needs numbers in a format software can act on. Morpho said agents can already be prompted to find the highest-yielding vault on Base, simulate a deposit of 10,000 United States dollar coin, build an unsigned transaction, and send it to a wallet for approval. (morpho.org) Morpho framed the launch around a fast-growing agent market. The company said more than 130,000 artificial intelligence agents have registered onchain identities since January, citing Dune data, and argued that lending is the next category those agents will enter after payments. (morpho.org) The pitch is not that agents replace the protocol’s existing interfaces; it is that they get a cleaner machine layer underneath them. Morpho still tells developers to add fallback mechanisms because its application programming interface can change and uptime is not guaranteed. (docs.morpho.org) DIA makes a similar case from the oracle side: lending systems need auditable values for collateral that may have no meaningful secondary market. Its site says it computes contract exchange rates for yield-bearing tokens, net asset value for vault positions, and reserve-backing ratios for collateral verification across more than 60 blockchains. (diadata.org) Put together, the setup is straightforward: Morpho exposes the lending rails, DIA supplies valuation data for harder-to-price assets, and agents get a machine-readable path into both. The remaining test is whether developers trust those tools enough to let software, not just people, make the first pass at lending decisions. (morpho.org) (diadata.org)
