How Nous Fi works.

What is Nous Fi?

Nous Fi is a tool that helps you do three things with your crypto — lend it, borrow against it, or stake it — without having to check ten different apps to find the best rate.

Normally, if you want to earn yield on your crypto, you'd have to visit Aave, then Compound, then Morpho, then a handful of others, compare rates by hand, and manually move your money whenever something better shows up. Nous Fi does that comparison for you, and (in a future version) can move your money for you automatically — but always inside limits you set yourself.

The two things working together

There are two separate systems behind every number and every suggestion you see in the app, and it matters that you understand the difference:

1. The AI looks at every available opportunity and picks the ones that best match what you asked for — how long you want to commit your money, how much risk you're okay with, what return you're hoping for. It also writes a plain-English explanation of why it picked what it picked.

2. The Risk Engine is a separate, much more boring piece of the system. It doesn't "think" — it just checks hard rules. Is this protocol trustworthy enough? Is the position too concentrated in one place? Would this borrow put you too close to liquidation? If the answer to any of those is "yes, that's too risky," the Risk Engine blocks it — and the AI has no way to override that block. The AI can only ever recommend things that already passed the Risk Engine's checks.

Think of the AI as a very well-informed assistant, and the Risk Engine as the strict rulebook that assistant is never allowed to break, no matter how good an opportunity looks.

Is this real? Can I use real money?

Not yet — this is a demo. Every number you see (balances, interest rates, positions, the AI's past decisions) is example data, not something pulled from a real blockchain in real time. Connecting your wallet works, but no version of the app currently takes your funds or moves anything on your behalf. You'll see an orange banner reminding you of this on every page inside the app.

This documentation still describes everything exactly as it works today, including the parts that are demo-only, so that when a live version does exist, nothing about how it behaves will be a surprise.

The five things you can do

Where to go next

• New here? Read The Dashboard to see what the home screen shows you.
• Want the AI to find something for you? Read Strategies.
• Want to browse and pick something yourself? Read Markets.
• Confused by a specific number (health factor, risk score, APY)? Check Risks Explained — every term is defined there.
• Something not covered? Check the FAQ at the end.

The Dashboard

This is the first thing you see after opening the app (/app). It's a summary of everything happening in your account — nothing here is something you configure, it's all a read-out of positions you've already opened elsewhere in the app.

The eight numbers at the top

The chart

A 30-day line showing your total portfolio value over time, so you can see at a glance whether it's been trending up or down.

Portfolio risk score

A single 0–100 number (lower is safer) summarizing how risky your whole portfolio is right now, averaged across every position you hold. Click "Open Risk Center" to see exactly what's driving that number position by position.

Protocol allocation / Asset allocation

Two bar lists showing where your money is — split by which protocol it's in (Aave, Sky, Convex, etc.) and which asset it's in (USDC, ETH, crvUSD, etc.). Useful for spotting if you're too concentrated in one place — the Risk Engine enforces limits on this (see Risks Explained), and these bars are how you can see your exposure yourself before it becomes a problem.

Upcoming maturities & repayments

A short list of what's coming up soon: positions that are about to reach the end of their term, and any borrowed amounts you'll need to repay. Each row shows how much time is left.

AI suggestion

A live example of what the AI would currently recommend under your "Safe Income" settings — protocol, asset, expected return after fees, and a short explanation of why it was picked over the alternatives. Click "Build a strategy" (top right of the page) to go set up your own criteria instead of using the default.

Recent automatic actions

A short feed of the most recent things the AI/automation system has done (or, in this demo, a record of example actions). Click "Full log" to see every action ever taken, with the complete reasoning behind each one — that's the AI Agent Log page.

Building a strategy

There are two ways to get the AI to find something for you: fill out the full form yourself (AI Strategy Builder), or pick one of four pre-filled presets (One-Click Strategies). Both end up doing the exact same thing under the hood — the presets are just the form filled in for you with sensible defaults.

AI Strategy Builder (/app/strategy-builder)

A form split into five sections. Every field on it directly controls what the AI is and isn't allowed to suggest.

Capital & horizon

• Capital — how much money (in dollars) this strategy is for.
• Investment horizon — how many days you're willing to have this money tied up.
• Allowed assets — which coins the AI is allowed to use. Click a chip to toggle it on/off (highlighted = allowed).

Return & risk targets

• Minimum APY — the lowest interest rate you'd accept. Anything below this is thrown out immediately.
• Target APY — what you're actually hoping for. The AI ranks options by how close they get to this number, not just by "biggest number wins."
• Max risk score — a 0–100 ceiling (0 = safest). Anything riskier than this number is rejected outright, no matter how good the return looks.
• Max drawdown — the most your position's value should be allowed to dip before you'd want out.

Protocol & chain preferences

• Allowed protocols — leave empty to allow all of them, or pick specific ones you trust.
• Blocked protocols — protocols you never want used, even if everything else about them looks fine.
• Allowed chains — which blockchains you're willing to use.
• Allow automatic bridging between chains — whether the strategy is allowed to move your funds from one chain to another to chase a better rate.

Borrowing & leverage

• Allow borrowing against collateral — off by default. Turn this on only if you want the strategy to be able to take out loans as part of chasing yield (see Risks Explained for why this needs extra caution).
• Max LTV ("loan-to-value") — the largest percentage of your collateral's value you're willing to borrow against.
• Min health factor — the closest to danger you're willing to let a borrow get before it should be paid back automatically. Higher = safer.

Automation & profit routing

• Automatically reinvest matured positions — when a position finishes its term, should the profit go straight back into a new position, or sit as free balance?
• Profit routing — where earnings go: back to your wallet, reinvested into lending or staking, held as stablecoins, or used to regularly buy a specific asset (DCA).
• Rebalance frequency — how often the strategy is allowed to re-check for a better opportunity and move your money.
• Max gas price — a ceiling on transaction cost, so the strategy won't move your funds if doing so is too expensive at that moment.
• Require my confirmation before every automatic action — if on, nothing happens without you approving it first, even if everything else about the move fits your rules.

After you click "Generate strategy"

You'll see:

• The recommended opportunity — protocol, asset, net APY, entry/exit cost, lock-up period, the top risks, a worst-case scenario, and specific reasons the position would later be closed or changed.
• Up to three alternatives that also qualified, ranked below it.
• How many opportunities were looked at and rejected, and why.
• "Run strategy in demo mode" — simulates starting the strategy. In this build, nothing real happens; it's there so you can see what confirming a strategy would feel like.

One-Click Strategies (/app/strategies)

Four ready-made versions of the form above, each already filled in with a different philosophy. Each card shows its own current top pick — this isn't the same recommendation repeated four times, each preset genuinely searches with different rules and can land on a completely different opportunity.

• Safe Income — only high-trust protocols, stablecoins or low-volatility assets, no borrowing, a conservative health factor. Priority: don't lose money.
• Max APY — the widest net across protocols and assets, clearly higher risk, shown transparently rather than hidden behind a big number. Requires your confirmation on every action.
• Min Risk — only the most audited, highest-TVL protocols. Willing to accept a lower return if it means a meaningfully lower risk score.
• Stablecoin Strategy — stablecoins only, spread across several protocols so no single stablecoin or platform holds too much of your money.

Click "Run this strategy" on any card to simulate starting it, same as the Strategy Builder's demo run.

Browsing markets: Opportunities, Lending, Borrowing, Staking

If you'd rather look at the raw list of what's available and pick something yourself instead of asking the AI, these four pages are for that. All four show the same underlying list — Lending, Borrowing, and Staking are just that same table pre-filtered to one category, so if you've read about one, you already understand the other three.

Opportunities Explorer (/app/opportunities)

The full, unfiltered list. Every row is one specific opportunity — a protocol paired with an asset, chain, and action.

• Tabs (All / Lend / Borrow / Stake) — filter the list by action type.
• Search box — type an asset symbol or protocol name to narrow the list.
• Sort dropdown — order by net APY, risk (lowest first), protocol TVL, or available liquidity.

What each column means

Click "Use" on any row to jump into the Strategy Builder with that opportunity as your starting point.

Lending (/app/lend), Borrowing (/app/borrow), Staking (/app/stake)

Exactly the table above, already filtered to one action, as dedicated pages — useful if lending is the only thing you care about and you don't want to touch a filter every time.

• Lending: supply an asset, earn interest from people borrowing it.
• Borrowing: put up collateral, draw a loan against it, pay interest on what you drew.
• Staking: lock an asset into a protocol for rewards — sometimes with a fixed lock-up period, sometimes flexible.

See Risks Explained for what can go wrong with each of these specifically — they don't all carry the same kind of risk.

Your positions, and everything about risk

Active Positions (/app/positions)

Every position you currently have open, in one table (or, on a phone, a stack of cards). For each one you can see the principal you put in, its current value, profit so far, APY, a progress bar showing how close it is to maturity, and — if it has a loan attached — its health factor. Click "Details" on any row to open the full picture for that one position.

Position Details (/app/positions/[id])

Everything about a single position:

• Principal / Current value / Accrued profit / APY — the basics.
• Maturity bar — how far along the position is, and the exact date it matures.
• Attached loan (if there is one) — the borrowed amount, its interest rate, LTV, health factor, and due date.
• Rewards accrued (if there's no loan) — any separate reward tokens this position has earned on top of its base rate.
• Risk assessment — a breakdown into six specific risk types for this exact position (see below for what each one means), plus a written worst-case scenario in plain English.
• AI activity on this position — every automatic decision the AI has made about this specific position, with its reasoning.
• Related transactions — the on-chain actions tied to it.
• "List on Marketplace" — only shown if this position is transferable (see the Marketplace page). If the underlying protocol doesn't support transferring ownership, you'll see a disabled "Not transferable" label instead, so it's always clear upfront whether early exit is even possible for that position.

Risk Center (/app/risk)

Two things live here: the rules that apply to everyone no matter what strategy they're using, and a breakdown of your own current positions.

Global risk limits

These are hard ceilings enforced by the Risk Engine for every single position on the platform — no strategy, however aggressive, is allowed to cross them:

Active risk alerts

Live warnings — for example, a stablecoin trading slightly off its peg — tied to specific positions you hold, so you know immediately if something needs attention.

Per-position risk breakdown

The same six-factor breakdown shown on individual Position Details pages, but for every position at once, so you can compare them side by side.

---

Risks explained

Every risk score you see in the app is broken into these categories. Here's what each one actually means, and which of your activities are exposed to it.

Smart contract risk — the code a protocol runs on could have a bug or be exploited by an attacker, potentially losing part or all of what's deposited in it. Applies to lending, borrowing, and staking alike. More audits and a longer track record generally (not certainly) mean lower risk here.

Liquidation risk — only applies if you've borrowed against collateral. If the value of your collateral drops (or the value of what you borrowed rises) enough, the protocol can automatically sell your collateral to cover the loan — sometimes at a worse price than you'd have liked, and sometimes with very little warning during a fast market move. This is why health factor and LTV limits exist.

Depeg risk — stablecoins are supposed to always equal $1 (or whatever they're pegged to), but during periods of market stress they can temporarily — or permanently — trade below that. If you're lending, staking, or holding a lot of one stablecoin, a depeg directly reduces your position's real value even though the token count doesn't change.

Liquidity risk — how easily you can get your money back out. A pool with low available liquidity can mean you're stuck waiting, or have to accept a worse price to exit quickly. Shown per-opportunity as the "Liquidity" column on the Markets pages.

Oracle risk — protocols need a reliable price feed to know what your collateral or position is currently worth. If that price feed (the "oracle") is wrong, delayed, or manipulated, it can trigger an incorrect liquidation or let someone else exploit the mispricing.

Bridge risk — only applies if a strategy is allowed to move funds between chains. Cross-chain bridges are a common target for exploits; funds in transit between chains carry risk independent of the protocols on either end.

No guaranteed returns — every APY shown anywhere in the app, including AI-projected returns, is an estimate based on current or historical rates. Rates change constantly based on supply and demand in each pool. Nothing in this app promises a specific future return.

Automation risk — if you let a strategy manage itself (auto-reinvest, automatic rebalancing, automatic loan repayment), you're trusting the system to act correctly and promptly. The Risk Engine's limits exist specifically to contain how much damage a mistake or a missed edge case could do, but they don't reduce that risk to zero.

Position Marketplace (/app/marketplace)

The problem this solves

Say you lent money for three weeks to earn a good rate, but after one week you need that cash back. Normally you'd either be stuck waiting, or you'd have to accept an early-withdrawal penalty, if the protocol even allows early withdrawal at all. The Marketplace is a second option: sell your position to someone else who's happy to wait out the remaining time and collect the final payout themselves.

Who's who

• Seller — someone who opened a position (lending, staking) and wants liquidity before it matures. They list it for sale at whatever price they choose.
• Buyer — someone with spare capital who's willing to wait. They pay the seller's asking price now, and in exchange take over the position: its full history, its remaining term, and the profit it will eventually pay out at maturity.

Nothing about the underlying position changes when it's sold — the same deposit keeps earning the same way. Only who is entitled to the payout changes.

Listing a position (as a seller)

Go to a position's Details page and click "List on Marketplace." This only appears if the position is transferable — some protocols' underlying tokens simply don't support ownership transfer, in which case you'll see a disabled "Not transferable" label instead and selling isn't possible for that specific position. If it is transferable, you'll set an asking price and publish the listing.

What every listing shows (as a buyer)

Click "Place bid" to make an offer on an open listing.

Should you buy or sell at a discount or a premium?

That's a judgment call the app doesn't make for you — it just shows you every number needed to decide. A position at a discount to its computed value can be a good deal if you're confident it'll pay out as expected; a seller might accept a discount simply because they value getting cash now more than waiting. There's no "right" price — it's a negotiation between two people who value time and liquidity differently.

Analytics, history, the AI's activity log, notifications, and settings

Portfolio Analytics (/app/analytics)

A deeper cut of the same portfolio data shown on the Dashboard, for when you want more than the summary:

• The same 30-day portfolio value chart.
• Profit by action type — how much of your total profit came from lending, how much from borrowing-related strategies, and how much from staking, shown separately.
• Allocation by protocol and allocation by asset — the same idea as the Dashboard's bars, with exact dollar amounts alongside the percentages.
• A short written note putting your blended APY next to your risk score, in plain language — e.g. pointing out which specific position is carrying most of your portfolio's risk budget.

Transaction History (/app/history)

Every on-chain action tied to your account, whether you triggered it yourself or the AI did it automatically on your behalf — deposits, withdrawals, borrows, repayments, reward claims, swaps. For each one you can see the amount, which chain it happened on, the gas cost, whether it confirmed successfully, the exact time, and its transaction hash.

AI Agent Log (/app/agent-log)

The complete record of every decision the AI/automation layer has made. This is the most transparent page in the app — nothing here is summarized away. Click any entry to expand it and see:

• Situation before — what was true right before the decision was made.
• Alternatives considered — the other options that were on the table, and why they weren't chosen.
• Reasoning — the actual explanation for the choice that was made.
• Expected benefit — what the decision was projected to gain.
• Risk before / after — exactly how the decision changed your risk score, in both directions (some decisions reduce risk, like repaying a loan on schedule).
• Related transactions — the on-chain actions tied to this decision.
• Outcome — what actually happened.
• Audit trail — a timestamped record of who did what: the AI proposing something, the Risk Engine approving or rejecting it, and the action executing. This is the literal proof that the Risk Engine is checked before anything happens, not after.

You'll also find entries here where the system rejected an opportunity — for example, something offering a very high return that didn't meet the minimum audit requirement. Those get logged too, so there's a record of what the AI considered and correctly turned down, not just the things it approved.

Notifications (/app/notifications)

Alerts about your account — position events, completed AI actions, and risk warnings (like a depeg watch), each marked info, warning, or critical. Click "Mark read" to dismiss one.

Settings (/app/settings)

• Wallet & session — shows your currently connected address, and confirms you're in demo mode (no custody, no real transactions).
• Default automation preferences — the same auto-reinvest, confirmation-required, profit-routing, and max-gas settings from the Strategy Builder, but set here as your default for any new strategy you create, so you don't have to re-enter them every time.
• Safety limits — a personal daily loss limit, on top of (never looser than) the platform-wide limits enforced in the Risk Center.

Frequently asked questions

Is my money safe? Right now, nothing is at risk — this build is a demo, and no real funds ever move. Once a live version exists, "safe" will depend on the same things it does for any DeFi product: the protocols it connects to, market conditions, and the limits you personally choose to set. Read Risks Explained before ever using a live version with real capital.

Do I need to connect a wallet to look around? No. You can browse every page, run the Strategy Builder, and read every number without connecting anything. A wallet is only needed to simulate taking an action (and, eventually, to actually take one).

What happens when I click "Run strategy in demo mode" or "Place bid"? In this build, it simulates the action locally (with a short loading animation) and shows you a confirmation — it doesn't send a real transaction or move real funds anywhere.

Why does the AI sometimes recommend something with a lower APY than another option on the same page? Because the AI isn't just chasing the biggest number — it's finding the closest match to your target APY while respecting your risk ceiling. A slightly lower rate at meaningfully lower risk can rank higher than a flashier number that would've pushed past your risk limit.

Can the AI ever ignore my limits if it thinks it knows better? No. This is enforced in how the system is built, not just a policy — the AI only ever gets to rank a list of opportunities that have already been filtered by the Risk Engine. There is no path in the code for the AI to reach an opportunity the Risk Engine rejected.

What's a "health factor," in one sentence? A number that tells you how close a borrow is to being automatically liquidated — above roughly 1.3–1.5 is generally considered a safer zone, and 1.0 means liquidation is imminent.

What does "risk score" actually measure? A single 0–100 number combining several underlying risks (smart contract, depeg, liquidity, liquidation, oracle, bridge — see Risks Explained) into one comparable figure, so you can rank very different opportunities against each other at a glance.

Why do some positions say "Not transferable" on the Marketplace button? Because the underlying protocol's token for that position doesn't support changing ownership. Rather than pretend that's possible, the app disables the option outright for those positions.

What's the difference between "Net APY" and "Gross APY"? Gross is the advertised headline rate. Net is what you'd actually keep after estimated entry/exit costs and fees — always the more honest number, and the one used for AI ranking and sorting.

Which blockchains does this actually work on? The current build is designed around Ethereum, Arbitrum, Base, Optimism, and Polygon. The site also references a much longer list of chains (BNB, Avalanche, Solana, Tron, Sui, Aptos, Sei, Hyperliquid, Sonic, Berachain, Mantle, Scroll, Linea, zkSync) as part of the long-term vision for unified liquidity across ecosystems — not all of them are wired up yet.

I found a number I don't understand that isn't explained here — what do I do? Every page's documentation above lists every field shown on it. If something's still unclear, it's worth treating that as a sign the product itself needs a clearer label — good feedback either way.