pitch.apis.accountants
The functions an accounting firm's systems call.
Document extraction, transaction classification, deterministic accounting math, close tracking — called mid-engagement by your systems and agents, at posted rates, with a receipt for every call. The licence, the opinion, the signature, and the client stay with the firm.
↓ scroll · arrow keys
Every close and every return stops in the middle for work that is neither the judgment nor the licence: pull the figures out of bank statements and invoices, put transactions where they belong, run the schedules, chase what the file is still missing. Today that work is staff hours plus a chain of point tools — each priced by seat or by negotiation, none of it callable at the moment the practice stack needs it, and none of it re-performable when a reviewer asks how a figure was produced.
That was tolerable when humans drove every step. It breaks now, for two reasons at once. The caller is increasingly a machine — a close checklist, a workflow engine, an agent acting for either — and a machine cannot sit through a seat-licence procurement; a price it cannot read at call time does not exist for it. And the human capacity the old model assumes is not being replaced.
The AICPA's own Trends reporting — the profession's pipeline series — documents multi-year declines in accounting graduates and CPA-exam candidates. The capacity story is the profession's own reporting, not a vendor pitch; per estate rule the figures stay in the source.
The contract is small enough to state in full. discover the catalog — typed schemas, posted rates, versions. describe one function's contract. estimate a call before making it — the estimate is binding for the quoted inputs. extract documents — bank statements, invoices, receipts, payroll reports — into structured, typed records. classify transactions against the firm's own chart of accounts, versioned and re-performable, with anything under the firm's confidence threshold routed to review instead of guessed. compute deterministic accounting math — depreciation and amortization schedules, accruals, sales-tax figures, reconciliation and trial-balance checks — versioned so the same inputs always reproduce the same figures. track what a close is missing, what is stale, what needs a chase. inspect the receipts.
// illustrative product mechanics, not results
POST /invoke
{
"function": "compute.depreciation@v7",
"program": "cas-close-2026",
"inputs": { "basisUsd": 48000, "method": "MACRS",
"convention": "half-year", "recoveryYears": 5 }
}
// every call returns a receipt:
{
"receipt": {
"function": "compute.depreciation@v7",
"program": "cas-close-2026",
"credential": "org:firm/systems/close-prod",
"rate": "cited from the public card",
"outcome": "ok"
}
}
Nothing is allowed to change silently underneath a close: the contract binds every function version to a published deprecation schedule, and every receipt to citing the exact version and the posted rate it settled at. Receipts are designed to be the one record a firm's reviewers and its own books both reconcile from — re-performance is the profession's standard of evidence, and the rail is built to meet it by construction.
The licence. The opinion. The signature. The client. This rail extracts, classifies, computes, and tracks; it does not express an attest opinion, sign or file a return, advise a taxpayer, or hold client money. Any step a statute reserves for a licensed person belongs to the firm's own licensed staff, and the rail is built so that routing is explicit.
The signature line is statutory: anyone who prepares or assists in preparing federal tax returns for compensation must hold a valid PTIN. The rail is designed never to hold one — the preparer of record is always the firm's human.
That boundary is the design, not a limitation. A rail that tried to absorb the licensed steps would become a competitor to its own customers; a rail that supplies functions into licensed engagements is infrastructure to all of them at once — including to firms who compete with each other. Neutrality is what the boundary buys.
Where the client's data goes is a contract term, not an appendix. Extraction and classification mean client books and tax return information cross into the rail mid-engagement — data whose use and disclosure federal law restricts with criminal teeth — so the rail is built to be read as the firm's service provider, never a second user of the data. Data is program-scoped like the credential that carries it, processed to execute the invoked function and produce its receipt, and never used to train models. Retention is bounded by the program's stated retention window, and subprocessors are disclosed on a published list before they touch a program's data — not discovered in an audit.
How the rail sits relative to IRC §7216 tax-return-information rules, Circular 230, and state accountancy confidentiality duties is ratified by counsel and published as part of the capability contract before any function serves live client data. The service-provider posture is a legal conclusion, not a design choice, and it is not claimed as fact here.
The control posture — access, retention, subprocessor management — is asserted as an audited fact when the SOC 2 Type I report is published, not before. Until then it is a design commitment, written into the contract where a firm's reviewers can hold the rail to it.
| name | unit | covers |
|---|---|---|
| Extraction | per invocation | bank statements · invoices · receipts · payroll reports, to structured typed records |
| Classification | per invocation | transactions to the firm-declared chart of accounts, versioned and re-performable |
| Deterministic math | per invocation | depreciation · amortization · accruals · sales tax · reconciliation and trial-balance checks |
| Close tracking | per invocation | missing · stale · needs-a-chase, across an engagement lifecycle |
Professional services is the home of the negotiated engagement letter, which is exactly why this rail refuses one. The caller is a system, so the price has to be readable at call time by the thing deciding whether to call. One card means the studio's own brands pay what a top-hundred firm pays — list price IS the arm's-length price, and third parties paying it are what makes that true. Every receipt cites the posted rate by contract, so a firm closes its own books against the same public card its reviewers audit agent spend against. Volume does not buy a different rate: on a mostly fixed-cost platform, volume improves our margin at the same card, and the card staying flat is what keeps the rail neutral between a two-partner shop and a national practice.
Per-invocation rates by function class. The structure is fixed — metered, classed by compute and latency profile, program-scoped; the figures publish when the card is live, and no figure is asserted before then.
Machine settlement over x402 / Mandate rails is the designed payment path. Until the first settlement clears on those rails, payment is by funded program balance drawn down per receipt.
Accounting demand is shaped like a calendar, and a metered rail does not get to hide from it: per-invocation revenue rides the profession's cadence by construction — compliance season concentrates extraction and computation into a few brutal months, while the monthly close runs as the year-round metronome.
Three properties are designed against the shape. Metered-with-no-minimums is season-shaped: when the season ends, a metered bill falls with volume by itself — there is no seat count to renegotiate in May, so the integration survives the trough that kills per-seat tooling. The catalog spans both sides of the calendar: season work is extraction and schedule math; the off-season is close tracking, reconciliation checks, and the chase list — the catalog grows by demand evidence, so it grows on whichever side the calendar is starving. The rail's own cost base doesn't ride the season: no licensed human supply sits on this brand's books — human accounting capacity is the supply side's business, not the rail's — so a seasonal trough compresses revenue, not viability.
Primary motion is B2A2B: our functions serve a firm through its systems and agents — the practice stack calling the rail mid-engagement on its own behalf. Secondary is A2A: external agents transacting with the rail directly, machine to machine, settlement included. The channel matches the motion — machine-readable discovery at the apex domain and position in the agent-discovery namespace, because a rail whose caller is a system is found the way systems look, not the way buyers browse.
A2A is the least proven motion in the estate and the largest one if it lands. It is claimed when the first external settlement clears, not before.
The functions on this rail — extraction, classification, deterministic math, close tracking — are precisely the work that migrates Human → Agentic → Generative → Code. The statutory floor in accounting sits with the licence-holder, which is the customer; the rail itself carries no licensed-supply cost, so its economics are software economics on a mostly fixed-cost platform, and margin improves with volume at an unchanged public card.
Calling-system count, invocation volume, and revenue. Not presentable until the reporting basis resolves; no figure appears in this deck before then.
This coordinate was not inherited, defaulted into, or picked up in a basket — it was acquired deliberately, as a pair, before a line of the product existed: the demand rail (this domain) and the supply-side door for human accountants on the same vertical, in the same acquisition run. The estate buys its coordinates before it builds on them, and the paper trail is public.
apis.accountants was registered 2026-07-26 via Porkbun — a standard-fee registration, not premium inventory, taken on a same-day retry after two failed availability checks, per the estate's acquisition log family (domains/data/apis-accountants-buy-results.json and rails-retry*.json — the failed checks and the successful registration each in their own run's log). The registration date and registrar are publicly verifiable in RDAP.
gigs.accountants — the supply-side pairing — was registered the same day, eighteen minutes apart in the same acquisition sweep, per the same log family (domains/data/rails-retry*.json). Demand rail and human door for one vertical, acquired together on purpose.
in the agent-discovery layer, accounting should resolve to apis.accountants the way automotive resolves to auto.dev — the domain is the shelf position for machine callers
a firm that has closed a year of client books against versioned receipts does not re-verify a new vendor for sport — the workpaper trail is the switching cost
posted rates let the rail serve competing firms simultaneously; a negotiated-price vendor structurally cannot be the neutral default for a whole profession
versioned functions reproduce the same figures for any reviewer at any later date — the property peer review standardizes on and then defends
Stated plainly: this deck precedes everything except the coordinate. apis.accountants is a G4 projection of the studio's G3 stack — the function runtime (functions.do), API layer (apis.do), and data layer (database.do) — and no product surface has shipped: no catalog, no card, no callable function. The zone is live and the apex today serves the estate's RESERVED register leaf — a placeholder filed on purpose, not a front door. Every amber below is deliberate: the deck is built before the catalog, and says so.
The apex serves today: the estate's own RESERVED register leaf — "apis.accountants · RESERVED", filed in the apis.finance family register — a placeholder posted on purpose. Serving is a liveness fact, not a product surface.
What the apex serves is a placeholder, not a product surface. The first product claim this brand can post is a real front door.
The catalog — typed schemas, posted rates, latency classes, versions — is the primary surface, designed so a system never needs an HTML page to find, price, or call a function. It is claimed when a machine can fetch it cold.
The rail is built to the Agent eXperience Protocol posture — discoverable, priceable, callable by machine. Conformance is claimed only when a cold conformance run is published.
The claim that matters: an external firm's production practice stack, calling under its own program-scoped credential against a funded budget, re-performing from receipts. It posts when it has happened, with the receipt discipline as the evidence.
The contract precedes the code: this record is the front door today — read it, then bring one close to scope against it.
If this was forwarded to you: apis.accountants is the demand rail into accounting — document extraction, transaction classification, deterministic accounting math, and close tracking, called mid-engagement by a firm's systems and agents at posted public rates, with a receipt for every call. The licence, the opinion, the signature, and the client stay with the firm; the rail supplies functions into licensed engagements and never substitutes for the licence. Every claim above carries its own state and evidence — including the many not yet earned; this brand is candidly pre-launch. If you run a firm's practice stack: read the contract. If you know who does: forward this.