Product Pitch — The Technologist (Why Now)
One-liner
TRD is a continuous, automated trust pipeline for private credit, built on the Sui stack: a TEE pulls the third-party data and publishes a daily aggregate while keeping the detail private, the aggregate is stored openly on an attested data layer (Walrus), the third-party API keys are held by Seal, and the facility runs as Move smart contracts on Sui. This page is a reference for that toolchain — what each component does, and why it only became buildable recently.
The hard part
A continuous trust pipeline for private credit has never really been built. Until now, validation meant human intermediaries doing periodic checks — an auditor once a year, a fund manager's quarterly review — slow, infrequent, and staffed by people who are not all incentivised to dig. The operating data exists; turning it into continuous, investor-readable evidence — automatically, and without piling new reporting burden on the borrower — needed automation and components that only recently matured. TRD is that pipeline, and much of the work is choosing the right components and wiring them together.
The stack — what each component does, and why
- Nautilus (TEEs) — keep detail private, aggregate the data. The TEE's job is to keep the detailed-level operating data private and produce the aggregated data, on a daily frequency. It does not monitor or compute the KPIs — the front end derives those itself from the aggregated daily data. Running the aggregation inside a Trusted Execution Environment makes it automated and attestable rather than a person compiling a spreadsheet, while the underlying detail never leaves the enclave in the clear. Nautilus made verifiable off-chain compute on Sui practical only recently.
- Walrus — the attested, public data layer. The daily aggregated data is stored on Walrus, the trusted, attested data layer with data-availability guarantees. It is public — anyone can see it — because open, attested aggregated data is the very base of trust at this level. (The detailed underlying data stays private in the enclave; only the aggregate is published.) The front end forms the KPIs from this public aggregate. Walrus is recent decentralized-storage infrastructure.
- Seal — custody of the third-party API keys. Seal holds and gates the API keys for the third-party web services the pipeline pulls from (Bolt, the LMS, the NTSA registry, the M-Pesa payment rail). It is programmable access control for those credentials — it does not encrypt the Walrus data, which is public by design. Programmable on-chain access control is itself a recent capability.
- Move on Sui — accounting, settlement & the facility. The accounting and the Open-Ended Adaptive Debt Facility run as Move smart contracts on Sui: deposits, drawdowns, repayments, the adaptive rate and its bounds, the cash buffer, payout epochs, and control states. One programmable layer where the data, the accounting, and the capital meet.
From snapshot to continuous
The conventional audit is a once-a-year snapshot — sample a few loans, trust the rest. TRD makes it continuous: the pipeline ingests independent sources (Bolt fleet activity, the lender's LMS, the NTSA registry, the M-Pesa payment rail), triangulates them, and refreshes the data daily. The auditor stops being a person with a clipboard and becomes an automated, always-on process — at no extra operational cost to the borrower.
Two innovations in one product
TRD is one product with two innovations: the trust pipeline above (surfaced as the Risk Monitor), and the Open-Ended Adaptive Debt Facility it feeds. The facility's adaptive rate is the elegant part — it prices risk continuously from pool utilisation and participation, bounded between borrower-set minimum and maximum rates, so the market can't run away from either side.
Why now
The components arriving together is the enabling shift; two things make it the moment. Recent RWA failures exposed the cost of weak, periodic, self-reported monitoring. And asset-level operating data is now reachable through APIs while stablecoin capital is hunting real-economy yield. The architecture is concrete, not a slide: consented ingestion, the TEE pipeline, Walrus storage, a browser verification console, on-chain accounting on Move, and the Risk Monitor front end — designed so the trust-experience surface comes first. Build sequencing lives in the roadmap.