The regulatory semantic layer for money transmitters
Quarterly state reporting, without the scramble.Every figure cited, lineage‑backed, reproducible years later.
TableFabric plugs read-only into the systems your money transmission business already runs — warehouse, ledger, processor — maps your data to what each state actually requires, and produces examiner-ready evidence packs for the NMLS MSB Call Report and the state reports alongside it.
- 0–0
- state money transmitter licences, held across roughly 600 companies
- 0 days
- after each quarter to file, in nearly every state that has adopted the form
- $0–0k+
- a year on licence maintenance alone, before any headcount
Why TableFabric
Built for money transmission, not adapted to it. One product, one industry: multi-state money transmitters and the reports their regulators actually read.
Your systems, left where they are
Your ledgers, processors and warehouses stay exactly where they are. TableFabric reads them in place through a read-only agent. Your raw rows never leave; nothing is ever written back.
snowflake.prod.finance.transactions.amountA regulatory semantic layer
Versioned, cited models translate what your columns hold into what a state examiner means by them — outstanding transmission obligations, average daily liability, per-state figures — and keep that translation alive as your schema and the rules change.
outstanding_transmission_obligation@0.2.0Pulled through to evidence
When an examiner asks where a number came from, the answer is already attached: the source columns, the calculation, the citation that requires it, and every caveat that qualifies it.
PI130 · verified against primary sourceLicensing platforms get you licensed. Catalogs give you tooling. Nobody owns the layer that proves, quarter after quarter, what your data means to a state regulator — with the lineage and citations to back it. That layer is the fabric in TableFabric.
The problem
Quarterly reporting is still stitched together by hand
Your compliance officer pulls engineering off product work. Numbers come out of a warehouse, a ledger, a processor and three spreadsheets. Someone reconciles them, someone else writes a fragile script — and when an examiner asks how a figure was produced, the evidence is reconstructed from memory.
It compounds. Every new state adds a filing, a bond calculation and another interpretation to keep straight. The cost of staying compliant grows faster than revenue does.
Outstanding obligations
A single figure that drives permissible investments and bond sizing — and that nobody can reproduce two quarters later.
Average daily liability
An average whose denominator is a regulatory choice, not an implementation detail.
State attribution
Sender, beneficiary or agent location produce different, individually plausible answers.
Exam evidence
The numbers are the easy part. Showing where they came from — lineage back to your source data — is what takes the week.
How it works
Watch a real column become a filed figure
Read-only from end to end, built for money transmitter data. Every label below is the product's own vocabulary — the qualified names, semantic roles and NMLS field ids are the real ones; the figures are from a run against a fictional licensee, shown to illustrate the shape of the output. Click any stage to explore it.
It holds SELECT and nothing else, and it attests the grants it actually observes — a claim your security team can check, not one we make on its behalf.
The honest part
Where the rules disagree, we say so
Some regulatory questions genuinely have more than one defensible answer. How a transaction is attributed to a state is one: NMLS explicitly leaves it to the filer and the state being reported to.
So we do not quietly pick. Each of those choices is an explicit election with a cited default, signed off by someone with authority to make the interpretation, and carried into every pack that depends on it. A silent assumption inside a bond calculation is the failure this product exists to eliminate.
And a figure that cannot be computed says why. It is never guessed.
Why not just a data catalog
A catalog hands you a blueprint. We ship the house — opinionated models for specific state regimes, and evidence in the shape an examiner expects.
| Category | What they do well | Where TableFabric sits |
|---|---|---|
| General data catalogs | Automated tagging, lineage, glossaries — a powerful blueprint for any domain. | The house, already built: state-specific semantic models and evidence in the shape an examiner actually asks for. |
| Licensing platforms | Get you licensed and keep renewals and filings on track. | Complementary. They handle the licence; we continuously prove what happened underneath it. |
| Spreadsheets and scripts | Whatever your team can rebuild by hand each quarter. | A system of record that survives a schema change, a staff change and an exam. |
Questions we get asked first
Does my data leave my environment?
No. The agent runs inside your network on a read-only, SELECT-only credential — it is structurally incapable of writing to your systems. Your raw rows never leave. The only things that cross to us are aggregates, at the grain a regulator asks for, and metadata such as column names and types.
Do you give legal advice?
No. We show what your data says and which requirement it maps to, with the citation and the lineage behind it. We do not assert that you are compliant, and the judgment stays with your compliance officer and your counsel.
Who is TableFabric for?
Multi-state money transmitters filing the NMLS MSB Call Report — and, where they apply, the state regimes that sit alongside it, such as California and New York.
Talk to us
We are working with a small number of design partners in the multi-state money transmission space — companies that have recently expanded, or have recently had a painful quarter.
Accounts are provisioned by hand. Tell us a little about your licences and we will get back to you.
We typically reply within one to two business days.