Frequently asked questions

Tessera lets you borrow USDC against tokenized stocks like tAAPL, tTSLA, and tSPY. An autonomous agent we call the Watcher watches every position around the clock to head off liquidations before they happen.

Two sides use it: lenders supply USDC and earn yield from borrower interest; borrowers post tokenized-stock collateral and draw USDC against it. Our tagline sums up the goal: "The safest place to borrow against tokenized stocks."

Tokenized stocks trade 24/7, but the stock market behind them doesn't. Prices gap overnight and over weekends when no one can react, and ordinary lending positions get liquidated in their sleep.

Tessera's answer is structural: conservative, gap-aware borrowing limits plus an AI risk layer that watches positions continuously and can act before a gap turns into a forced sale.

You supply USDC to the vault. Borrowers pay interest on what they draw, and that interest accrues to lenders in proportion to what they've supplied. You can withdraw your USDC plus accrued interest subject to available liquidity in the pool.

Yield is variable and not guaranteed — it depends on how much is borrowed and at what rate. A share of borrower interest (the reserve factor) is set aside before yield is paid; see the fees section.

Deposit a supported tokenized stock as collateral, then borrow USDC up to that asset's maximum loan-to-value. You repay whenever you like; interest accrues on the outstanding balance until you do.

1. Deposit collateral (tAAPL, tTSLA, or tSPY).
2. Borrow USDC up to the asset's max LTV.
3. Stay above a health factor of 1.0, or let the Watcher help you (optional).
4. Repay USDC any time to free your collateral.

There is a minimum debt of 100 USDC per position to avoid uneconomical dust loans.

The health factor (HF) is a single number that tells you how safe your loan is. It's your collateral value, risk-weighted by each asset's liquidation threshold, divided by your debt.

1.0 is the liquidation line. Above 1.0 you're safe; the higher, the more buffer. Below 1.0 your position can be liquidated. The Watcher's job is to keep you comfortably above that line.

Three tokenized stocks today, each with conservative, gap-aware limits set on-chain. The max LTV is how much you can borrow against the asset; the liquidation threshold is the point at which the position becomes liquidatable.

These limits are intentionally lower than typical lending markets, precisely because tokenized stocks can gap when the underlying market is closed.

The core is built on Arbitrum Stylus (Rust): a vault contract, a PriceGuard oracle-policy contract, and a read-only Lens for data. A TypeScript agent runs the Watcher off-chain but can only act through the vault's permissioned, capped entrypoints.

It's open source. You can read the code at github.com/Ritik200238/tessera ↗.

It watches every borrower's health factor around the clock, re-checking roughly every 10 seconds. When a position drifts toward danger, it sends a plain-English alert so you know before things get tight.

If you've opted in, it can also act — auto-repaying part of your debt to head off a liquidation. What it can and can't do is spelled out below.

Only in one narrow, user-authorized way: it can reduce your own debt using USDC you pre-approved, and nothing else. It never custodies your funds and can never move them to anyone.

No — a model never decides whether to move money. Every financial decision is made by a deterministic core: fixed rules that produce the same output for the same inputs, with no model in the loop.

A language model is used only to write the human-readable copy — turning a decision the deterministic core already made into a clear alert you can understand. This split is deliberate: the part that touches your money is auditable and predictable.

Revoke the USDC allowance you granted. That's the single kill switch — once revoked, the Watcher physically cannot touch your position, because its only power was that allowance.

There is also an on-chain admin fail-safe so the protocol can halt the agent if something goes wrong. Either path stops it immediately.

The agent stamps an on-chain heartbeat as it runs. A permissionless backstop is built and tested so that, once enabled, anyone can step in and liquidate if the agent goes silent past the configured delay — so the protocol doesn't depend on our agent staying up. On testnet the delay is set to 0 (backstop off, liquidation is agent-only); turning it on is an explicit mainnet gate.

Your funds are never trapped by the agent being offline; normal repay and withdraw functions remain available directly on the vault.

Opting in means setting a USDC allowance the Watcher can draw on solely to repay your debt. You set the amount, you can change or revoke it any time, and the on-chain caps still apply on top of it.

It does not give the agent custody, withdrawal rights, or any ability to act outside reducing your own debt.

This is the core risk Tessera is built around. Conservative LTVs and liquidation thresholds leave a buffer for exactly this, and the Watcher tries to repay into the gap. But a large enough gap can still cross the liquidation line before anything can react.

The honest takeaway: the design lowers this risk meaningfully, it does not remove it. Borrow with room to spare.

If your health factor falls below 1.0, a liquidator can repay part of your debt and receive your collateral at a discount. The base liquidator bonus is 5% (it ramps with depth). The close factor is 50%, meaning at most half a position is liquidated in one step — except below a health factor of 0.95, where a full close is allowed.

Liquidation is how the pool stays solvent for lenders. The buffers and the Watcher exist to make it rare, not impossible.

A dedicated PriceGuard contract sits between the price feed and the vault, enforcing sanity checks and policy on prices before they're used — so a single bad or stale print doesn't immediately drive liquidations.

We will not present an audit, or any security claim, that we don't actually have.

As with any DeFi protocol, the contracts could contain bugs, the interest-rate or liquidation logic could behave unexpectedly under stress, and an unaudited codebase carries real risk. The interest-rate curve is capped (a hard ceiling), and risk parameters are conservative, but no design is risk-free.

Because this is testnet with no real funds, today's risk is to test assets only. That changes at mainnet, which is why audit and a real price feed are gates.

In principle, yes — in any lending protocol, extreme market moves or bad debt can leave the pool short. The reserve factor builds a first-loss reserve to absorb shortfalls first, and conservative limits reduce the chance of bad debt, but there is no insurance and no guarantee of repayment.

There's no separate platform fee on deposits or withdrawals. The protocol's only economic take is the reserve factor: 15% of borrow interest, which funds protocol revenue and a first-loss reserve. Lenders earn the remaining interest.

Borrowers also pay a liquidator a 5% base bonus only if they get liquidated; that's an incentive paid to liquidators, not a fee Tessera collects.

From the reserve factor — the 15% share of borrower interest described above. It's transparent, on-chain, and tied directly to real usage of the protocol. Revenue grows only if people actually borrow.

We deliberately do not monetize through a token sale, because token incentives tend to distort a protocol away from the users it's meant to serve.

These restrictions apply now and will carry into any mainnet launch.

Tessera is founder-led and built in the open. Team bios are being finalized; in the meantime the work speaks for itself — the code, parameters, and disclosures are all public.

You can follow the build and read the code at github.com/Ritik200238/tessera ↗.

It's real and deployed — the vault, PriceGuard, Lens, and the Watcher all run live on Robinhood Chain (chain 46630), an Arbitrum Orbit L2 — but on testnet, not mainnet.