DeFi has never been short on data. The problem is what happens to it. A curator can publish a detailed risk report on every asset they touch, and still watch the analytics show roughly five clicks a month against a strategy that hundreds of thousands of people are allocated to. The best risk intelligence in the world does very little if it lives in a dashboard nobody opens at the moment a decision gets made. Newton and Vaults.fyi exist to close that gap, pairing the most accurate vault data in DeFi with the ability to act on it automatically, before a transaction settles. Here is how the pieces fit together, and why it matters whether you are an institution allocating capital or an individual earning yield.

First, what is a vault?

A vault is one of the simplest ways to put capital to work onchain: you deposit assets, and yield comes back. The concept traces to the ERC-4626 standard, which did for onchain fund vehicles what ERC-20 did for tokens, defining the basics of how to price shares, value assets, and handle deposits and withdrawals. Behind that simplicity sits a curator, the party who decides how the vault's assets are allocated, and allocators, the institutional or individual capital that sets the mandate for how far a curator can go. The word "vault" has since become a catch-all, stretched to describe almost any DeFi product, from tightly constrained Morpho and Euler vaults to open-ended structures where a manager can move funds to an exchange, a custodian, or an OTC options desk. That range is exactly why the health of a given vault matters so much, and why it is so hard to judge from the outside.

Newton: an authorization layer for onchain transactions

Think of Newton like the Visa network. Before a card payment settles, there is a real-time check for fraud and risk; only once it passes does the transaction go through. Newton does the same for onchain transactions, vetting and authorizing them before they settle. Where a smart contract can only enforce static rules (a supply cap, an approved address), Newton policies can check live data, such as a vault's health score dropping or an allocation drifting off mandate, and block the action if it fails. Those policies live in open, composable packs that a curator can plug into their wallet, with every check recorded on the Newton Explorer so depositors and issuers can confirm the rules exist and are being enforced. Newton's mainnet beta is live, and VaultKit, built by Magic Labs, prepackages these policies so integration is close to effortless.

Vaults.fyi: API for onchain finance

Vaults.fyi began as a technical team's rebalancing experiment and grew into a single API and interface for anyone who wants to integrate DeFi or read the data behind it. Today it covers market data, call data, and position and P&L data for roughly 100 protocols across more than 20 networks, the majority of onchain TVL, and tracks over a thousand vaults. Its foundation is accurate APY. Rather than trust each protocol's self-reported numbers, Vaults.fyi fetches the share price of every product it integrates every hour and recomputes the yield itself, which also surfaces the negative APYs that few front ends will show on their own. On top of that sits a reputation score, a zero-to-100 measure built today from five factors: TVL over time in the pool, TVL over time in the protocol, underlying liquidity for the LP token, the network, and the distribution of holders. More signals are on the way, including paused withdrawals and maximum drawdowns, two of the risks that tend to go unnoticed until it is too late.

How it all comes together

Vaults.fyi supplies rigorous, real-time data and scoring; Newton turns it into rules that cannot be bypassed. The integration runs deeper than a single number. A user building a strategy can set simple, precise guardrails: never allocate to a vault with less than a million dollars of TVL, or below a 4% APY, and gate allocations on the reputation score. Because the score updates live, so does the enforcement. When vaults.fyi flags a vault, the change reaches Newton immediately. During the recent Kelp incident, when Aave and a range of downstream vaults were affected, vaults.fyi added manual flags that rapidly cut the impacted scores. An allocator with a predefined list of fifty vaults would have seen the ten affected ones drop below threshold and fall out of their allocation framework automatically, with no dashboard-watching required. The market moves from reactive, tracing what went wrong after the damage is done, to proactive, preventing the bad transaction from settling in the first place.

A toolkit for a new kind of allocator

Enforcement also opens a door that monitoring alone cannot. Beyond curators raising capital, there is a growing class of prosumer users who want to assemble their own allocation framework: pick the specific markets they like, skip the ones they do not, and manage it over time. Newton can act as that allocation toolkit. Through a smart account with session keys, a user can program the rules once, add and remove vaults as they go, and let the authorization layer handle the rest, with vaults.fyi's endpoints feeding it live signals. For existing allocators who have stitched together four or five tools to build a janky version of this in-house, it replaces the duct tape with a standard.

Where this is heading

Put the pieces together, accurate data, independent scoring, and an enforcement layer that acts at the moment of truth, and the guardrails stop being something a person has to read and remember. Adoption is widening, definitions are firming up, and large institutions are expected to launch their own vault standards. As bots and eventually AI agents take over rebalancing and routine money movement, the data has to enforce itself, because no one will be reading the report at machine speed. That is the layer Newton and Vaults.fyi are building together.

See it on your own vault

Want to see Vaults.fyi-powered risk scores gating real vault actions through VaultKit, with no changes to your vault? Schedule your demo.