Run the facility on live data, not last month's tape

Eligibility, borrowing base, covenants, drawdowns, waterfalls and lender packs, all computed from the records that created the assets. Your covenant position is a live view, and every certificate re-runs byte for byte months later.

Continuouscovenant evaluation
No tapesread from source systems
Freecounterparty seats
Covenant monitor · WH-Aevaluated 41s ago
Advance rate cap85.0 / 85%At capSector concentration28.5 / 30%WatchTop-10 obligors11.2 / 15%Pass30+ DPD delinquency3.4 / 4%WatchWeighted average tenor96 / 120dPassOvercollateralisation113 / 108%Pass
10 pass2 watch0 breach12 of 12 machine-readable

Illustrative interface. The same definitions render your lender's compliance view, so the two sides cannot disagree on a formula.

Capabilities

Institutional-grade capital orchestration, controlled at source

Compliance is enforced where assets are created rather than reconciled after the fact, and cost is optimised from the first drawdown rather than the first audit.

Collateral quality testing

Asset-level eligibility at pledge and continuously thereafter, with exclusive-pledge checks enforced at the ledger across every facility.

Transparency to lenders, and fraud prevented at the source

Covenant monitoring

Financial, portfolio and structural covenants held as executable rules, evaluated on every engine run with amber and red thresholds.

Breaches are seen forming, not discovered

Borrowing base optimisation

Allocates eligible assets across facilities and tranches against an objective you choose: maximum base, lowest cost of funds, or most headroom.

Better balance-sheet management, measured in bps

Drawdowns and waterfalls

Multi-tranche allocation, advance requests, collections sweeps and payment waterfalls executed from the rules you modelled, not a separate spreadsheet.

Modelled and paid are the same rules

Lender pack generation

Borrowing base certificates, compliance certificates, monthly servicer reports, stratifications and variance reports, generated on schedule.

Reporting stops being a person's job

Scenario testing

Model an amendment, a new pool, a stress case or a proposed draw against the live constraint set before anything is committed.

Decisions taken with the answer in hand

The build

Pool to base, with every reduction attributed to a rule

Two kinds of reduction, kept separate: hard exclusions where an asset fails an eligibility rule, and value trims where excess over a concentration, tenor or FX limit is haircut without removing the asset. Trimmed value is exactly where the optimiser finds room.

Pledged collateral pool$141.2MHard exclusions, 264 assets−$4.1MValue trims, concentration and tenor and FX−$4.4MEligible collateral, 3,918 assets$132.7MBlended advance rate× 85.0%Borrowing base$112.8M
Drawn $96.4MHeadroom $16.4MCommitment $180M
Optimiser readout

baseline pool: $132.7M eligible at 85.0% advance optimised pool: $138.1M at 85.5%, delta +47 bps

Same shape, every time, so the number can be checked rather than trusted. The proposal is an explainable set of asset-level moves, and nothing applies without four-eyes approval.

See the animated build

Illustrative build for a $180M senior warehouse. Figures are examples, not client data.

Counterparty portal

Your lender reads the same number you do

Lender analysts, auditors, trustees and backup servicers get read-only access to the certificates, covenant tests, pool explorer and data room, whitelabelled to your brand. Their seats cost nothing, because a lender who can self-serve stops asking you for spreadsheets.

The result is the phrase the platform is built around: one calculation, two audiences.

Borrowing base certificate

Pinned to the ledger cursor it was produced from, so it re-runs identically on request.

Covenant compliance

Every test with its clause reference, current value, headroom and history, including cures and waivers.

Pool explorer

Loan-level visibility into the collateral supporting the facility, filtered and stratified on demand.

Reports and data room

Servicer reports, executed documents and an append-only audit trail that diligence can read without your help.

One layer down

How the engine actually works

Selecting a borrowing base is not a prediction problem. Covenants are constraints, and because several of them are measured at portfolio level, every asset in the pool contributes constraint rows of its own. Given an objective, the question is which eligible subset satisfies all of them at once.

Feesable solves that with a deterministic linear program. The same ledger state returns the same allocation, every time, which is the property auditors and credit committees care about. Covenant compliance should not be left to probability.

Constraints compiled4,214from 12 covenants and 6 eligibility rules
Decision variables3,918eligible assets in the pool
Solve time41son demand, nightly, and after material events
ReplayExactimmutable inputs plus a versioned rule

Machine learning has a place in the platform, and we are explicit about where: anomaly review on the loan tape and consistency checks on supporting documents. It is never in the calculation path.

Bring your facility document and a sample tape

We will digitise the covenants, run your pool through the engine, and show you your own borrowing base rather than a demo dataset.