Fund administration software for people who sign the books

An administrator's problem is not one fund. It is forty of them, each with its own agreement, its own calendar and its own idea of what a management fee is.

Solutions · roughly 6 minutes

Forty funds, forty agreements

The administrator's difficulty is variety. One client's management fee is charged on committed capital during the investment period and on invested capital after it. Another's is on net asset value with a most-favoured-nation clause. A third caps organizational expenses and makes the manager bear the excess. A platform with a house convention makes all of them somebody's manual adjustment.

Fee structures here are composable rather than parameterized: tiered schedules, modifiers, offsets, expense caps and four-way settlement, configured per fund. Terms can be read out of the limited partnership agreement into a prefilled builder, with the clause attached, and a human confirms every one before it becomes a fund term.

A close that actually closes

Period close is a state machine, per fund and per period: open, soft closed, hard closed. A hard-closed period freezes writes and its trial balance is captured as an artifact, so a statement produced a year later still ties to what was reported at the time. Guards are not advisory: a fund with negative capital or a pending crystallization is reported and skipped rather than forced through.

Corrections stay possible without pretending they did not happen. Anything in an open period is editable; a sealed period takes a graded, audited revision. The distinction matters because the alternative, in practice, is an administrator keeping a private list of what the system says versus what is true.

A Statement of Changes in Partners' Capital, from beginning capital through contributions, net income and fees to ending capital.
Ending capital ties to the dashboard NAV to the cent. Not because a reconciliation ran, but because both are reads of the same ledger.
A gross-to-net fee disclosure with total gross, adjustments and net, and per-account management fee accruals beneath.
Gross to net, per accrual, per client. Each posted fee keeps the locked step sequence it was calculated with - gross, then offsets, waivers and rebates, then net - read back from that snapshot rather than recomputed later against terms that may since have changed.

Breaks with somewhere to go

A reconciliation report that nobody works is a compliance artifact, not a control. Breaks here are records with a lifecycle: detected, assigned, explained, resolved, with the detectors that found them stated and an aged worklist rather than a monthly PDF.

A break detail panel showing expected, actual and difference, a note field, and buttons to acknowledge, resolve or accept the difference.
Expected, actual, difference, and a decision with a name on it. Resolve and accept are deliberately different: resolved means the numbers now agree, accepted means they still differ and somebody decided that is correct. Accepting requires a note, because "they differ and that is fine" is only institutional memory if it says why.

Every screen runs on the public API

There is no private back channel. The product's own interface is built on exactly the API a client gets, which is the only real guarantee that the API is complete. Custodians, banks, portals, data vendors and a client's own warehouse all integrate against the same surface, and the systems that will never have an API get bulk CSV loaders instead.

A cash movements blotter with settled movements, each portfolio transaction paired with its GL-derived counterpart.
Cash movements, with their sources. Bulk selection and row actions across the grids, because an administrator's day is spent on many rows at once rather than one.

Related reading

Currently in private development

Built for the people who have to sign the books rather than only read them. If this is your problem, we would like to hear how you have it.

Request early access