Parametric is the AI-powered risk engine for Real-World Asset collateral in DeFi. We convert messy off-chain data into transparent, dynamic borrowing parameters — so RWAs can scale safely.
Built for issuers, lenders, and protocols that want institutional-grade guardrails — without sacrificing composability.
Tokenizing an asset is easy. Making it safe collateral is the hard part. Most RWA collateral fails because the real world is dynamic — and parameters aren’t.
Parametric outputs a Collateral Score plus a full parameter set: LTV, haircut, borrow caps, and liquidation buffers — updated as conditions change.
Safety-first defaults: when uncertainty rises, the system tightens borrowing automatically. If critical feeds go stale, circuit breakers reduce exposure before losses compound.
“Explainable” isn’t optional. Each score includes human-readable drivers (data freshness, performance drift, anomaly flags, liquidity assumptions).
Parametric transforms off-chain signals into on-chain parameters that protocols can consume like any other primitive.
Cashflow history, repayment performance, asset metadata, jurisdiction & servicing details — normalized into a clean interface for vaults and protocols.
Models estimate default likelihood and loss severity, while anomaly detection flags duplication, circular flows, and behavior drift. Model disagreement triggers safety bias.
Borrowing tightens when risk rises, expands when confidence improves, and pauses when data becomes unreliable. Circuit breakers prevent silent systemic exposure.
Every module is built to reduce fragility: detect issues early, tighten exposure automatically, and keep outputs composable.
Haircuts adjust based on performance, data freshness, and volatility — not quarterly opinions.
Different risk tiers earn different borrowing capacity. Safer pools get better capital efficiency.
If feeds go stale, parameters tighten automatically. Stale data can’t silently become systemic risk.
Pattern detection flags suspicious repeats, circular flows, and servicer anomalies early.
If uncertainty spikes or models disagree, the system defaults toward safety-first parameters.
Ship faster with curated templates for common RWA types — then customize with governance.
Parameters ship as a primitive: plug into vaults, risk modules, and monitoring pipelines — no bespoke dashboards required.
No hype, no hand-waving. The UI makes it obvious what the model knows — and what it doesn’t.
“We finally stopped treating RWAs like static collateral. The staleness tightening alone changed how we think about exposure.”
“The score is useful, but the explainers are the product. Reviewers can see drivers instantly — and challenge assumptions.”
“Circuit breakers + caps turn ‘unknown unknowns’ into bounded risk. It feels like institutional guardrails on-chain.”
Parametric treats risk as a system, not a spreadsheet. Guardrails are first-class: caps, brakes, and staleness logic are built in.
PARA is designed for protocol participation and risk alignment — not as a promise of profits.
A phased path from MVP risk primitives to institutional-grade underwriting rails.
Data schemas, vault interface, parameter format, monitoring & staleness logic.
First RWA templates (e.g., invoices / leases), dynamic haircuts, basic circuit breakers.
Multi-oracle support, dispute flows, expanded templates, integrations with lending protocols.
Advanced stress regimes, enhanced explainability, and compliance-friendly extensions — without breaking composability.
Short answers. No fluff. Built for skeptics.
Any asset with reliable performance data and enforceable claims can be modeled. We start with categories where servicing and performance are measurable (e.g., invoices, leases, short-duration receivables), then expand.
No. Parametric consumes oracle/data feeds and outputs risk-aware parameters: scores, LTVs, haircuts, caps, and safety controls.
Parameters tighten automatically. If critical feeds go stale or anomalies exceed thresholds, circuit breakers reduce borrowing or pause it.
We design for adversaries: anomaly detection, multi-source validation, caps, and governance oversight. The goal is resilience — not wishful thinking.
PARA is used for governance participation and risk alignment (staking/backstop). Integrations can start permissionless, with advanced roles gated by stake and reputation.
Parametric is building the risk infrastructure that lets RWAs scale responsibly: dynamic parameters, transparent scoring, and automated guardrails.