A host's money, in columns
Hosting looks simple from the outside: guests pay, money arrives, you clean and repeat. The bank account tells a different story. Payouts land in batches that match no single booking, arrive days after check-in, and are already net of a fee you never see charged. Around them sit cleaners, laundry, restocking, utilities that swing with occupancy, insurance, and in most cases a mortgage.
All of it is printed in a PDF that no spreadsheet can read. FlowParse converts the statement with AI — every transaction with its date, description, signed amount and running balance — and checks the arithmetic against the balance the bank printed, so nothing is silently missing before you start.
What comes out is what short-let bookkeeping actually needs: a sortable table of what each property earned, what it cost to run, and what was left — from any bank, with no template to configure.
The payout is not the booking
This is the fact that makes short-let accounts harder than they look. A guest pays one amount. You receive a different, smaller amount, at a different time, sometimes bundled with other reservations, and the difference is a host service fee that is deducted before the money moves rather than charged to you afterwards.
So the number in your bank is net revenue, not gross revenue. If you total your payouts and call it income, you have understated both your turnover and your costs by the same amount — which quietly matters, because the fee is a deductible expense and because turnover is what tax thresholds and lender affordability tests are measured against.
Converting the statement gives you the banked side precisely: every payout, dated, with the platform identifiable from the description and the amount as a real number. Sitting next to the platform's own earnings report, the gross-to-net bridge stops being an assumption and becomes a calculation.
Why short-let is not the same as letting
If you let a property on a long tenancy, one predictable rent arrives from one person on roughly the same day each month, and the bookkeeping is close to trivial. That is a different job, and it has its own page: bank statement conversion for landlords.
Short-let is the opposite shape. Income arrives many times a month from a platform rather than from the person who stayed, already net of fees, with cleaning and restocking costs that recur on every turnover rather than once a tenancy. Occupancy drives your utility bills. A bad month is a calendar problem, not an arrears problem.
Both jobs use the same engine, but the questions differ. A landlord asks whether the rent arrived. A host asks what the payout actually contained, what the turnover cost, and whether the property earned more than it consumed this season.
What FlowParse pulls from a host's statement
Extraction is by meaning rather than by template, so it works with whichever bank holds your payouts: every transaction's date and value date, the full description including the platform reference or payout ID, the signed amount, the running balance, and the counterparty where the statement names one.
Descriptions carry the weight here. A payout reference is often the only thing distinguishing one platform's deposit from another's, and descriptions that wrap across several lines are joined back into a single field rather than truncated, so the reference you need is still there when you go to match it.
Every amount exports as a typed number with the correct sign, so payouts and costs total in a single formula rather than needing to be cleaned up first.
How to convert a host bank statement
Upload the statements
Drop in one or many PDF statements from any bank — digital or scanned.
Let AI extract them
Every transaction is read with its date, description, signed amount and balance.
Check the balance
Opening balance plus transactions is checked against the closing balance the bank printed.
Export and categorise
Download clean Excel or CSV — or an accounting-ready file for QuickBooks or Xero.
Cleaners, laundry and the cost of a changeover
Every departure triggers a cost: the cleaner, the laundry, the consumables that get replaced whether or not anyone used them. Guests are often charged a cleaning fee, which arrives inside the payout, while the cleaner is paid separately out of the account — so the two halves of the same event never sit next to each other.
That is how hosts end up subsidising their own changeovers without realising. If your cleaning fee has not moved in two years and your cleaner's rate has, the gap is real money, and it is invisible until both sides are numbers in the same table.
As structured rows, cleaner payments become a dated series you can total per property and per month, and the cost per changeover — the single most useful number in short-let economics — becomes something you can compute rather than estimate.
Tourist and occupancy tax: who actually remits it
In many cities the platform collects an occupancy or tourist tax from the guest and remits it directly, so it never touches your account and is not your income. In others, the platform collects nothing and the obligation is entirely yours, payable to the municipality on a schedule that has nothing to do with your payout schedule.
The two situations look almost identical from inside a payout, and getting them the wrong way round means either paying a tax twice or not paying one at all. Which applies depends on your city and your platform, and it changes.
FlowParse does not determine that for you and does not calculate the tax. What it does is show you the bank truth: if a tax is arriving in your account, it is a row; if you are remitting it, that is a row too. Those rows are the evidence your accountant or your municipality will ask for, and they are the thing that tells you which regime you are actually in.
The honest boundary: we do the bank side
This is worth stating plainly, because the gig economy is full of tools that overclaim. FlowParse does not connect to Airbnb, Booking.com or Vrbo. There is no integration, no API, no login to your host account — and we are not going to pretend otherwise.
What we convert is the statement your bank produced. The platform's own earnings or transaction report is the other half of the picture, and you download that from the platform. The value is in having both as clean data so they can be compared, because that comparison is where discrepancies actually surface.
We also do not issue tax forms, do not file returns, and do not tell you what is deductible in your country. We produce accurate, balance-checked, categorised data. What it means for your tax position is a conversation with your accountant — and this gives them something worth reading.
Two listings, one bank account
Hosts almost always start with one property and end up with several, and almost nobody opens a separate account for the second. So a single statement carries the payouts, cleaners, utilities and mortgage of two or three properties at once, mixed together in date order.
Untangling that by eye is the reason most hosts have no idea which property is actually the profitable one. The busy listing with the constant turnover and the expensive cleaner is not automatically better than the quiet one with the high nightly rate.
Once transactions are structured rows, attribution is a filter: payouts by listing reference, cleaner payments by property, utilities by supplier and address. Profit per property becomes a pivot table rather than a suspicion — and that is the number that decides whether listing number three is a good idea.
Utilities, supplies, insurance and the mortgage
Short-let costs behave differently from residential ones. Utilities move with occupancy rather than sitting flat, because guests do not economise on heating. Consumables run continuously. Insurance is a specialist short-let policy rather than a standard landlord one, and it costs accordingly.
The mortgage deserves particular care. The payment leaving your account is part interest and part capital repayment, and only one of those is a cost. Expensing the whole payment overstates your costs and understates your equity — a mistake that is easy to make and awkward to unwind.
In a structured statement each of these is a category rather than a guess. Categorising transactions once and reusing the rules month after month turns a year of bookkeeping into a routine that takes an evening.
The season, the cash and the quiet months
Short-let income is seasonal in a way that long-let income is not. A property can earn most of its annual revenue in four months and still owe a mortgage in the other eight, which means a profitable year and a difficult February are entirely compatible.
That rhythm is invisible in an annual total and obvious in a monthly series. Money in by month, money out by month, and the balance that results is the picture that tells you how much of the summer has to be carried into the winter.
It is also the picture that makes a refurbishment decision rational. Doing the work in the quiet season costs you little revenue and a lot of cash — and knowing which of those you are short of is precisely what the bank data shows.
Co-hosts, splits and shared properties
Plenty of hosting arrangements involve someone else: a co-host taking a percentage for managing turnovers and guest communication, a family member who owns half the property, a management company taking a cut before the money even reaches you.
Those splits create an obligation that lives entirely outside the platform. The payout arrives whole, and then a share of it has to leave — on a schedule and a percentage that only exists in an agreement between two people.
Structured payouts make that arithmetic auditable. The share paid out is a dated row against the payouts it was calculated from, which is exactly what you want when a co-host asks how their number was reached, and exactly what you need if the relationship ever ends badly.
Airbnb, Booking.com, Vrbo and direct bookings
Most established hosts run more than one channel, because each reaches a different guest and each charges differently. Airbnb takes a host fee. Booking.com typically bills commission separately, which lands as its own outgoing charge rather than being netted off. Direct bookings arrive gross, via a payment processor that takes its own cut.
Three channels, three completely different money shapes, all landing in one account. Comparing them fairly is impossible while they are text in a PDF, and it matters enormously, because the channel that fills your calendar is not necessarily the channel that pays you best.
Converted and categorised, each channel becomes a column: revenue net of its own cost structure, per booking and per month. That is the comparison that tells you where to push your availability next season.
Reconciling payouts against the platform report
The check every host should do and almost none do: does what the platform says it sent match what the bank says arrived? Usually yes. Not always. Payouts fail and get reissued, a refund gets deducted from a later batch, a resolution-centre adjustment appears with no obvious link to the booking it relates to.
None of that is detectable while your payouts are lines in a PDF. It becomes detectable the moment they are dated rows with amounts sitting beside the platform's own report.
Reconciliation here is not an accounting formality. It is how you find the payout that never landed, and the only realistic window for doing something about it is before the year is over.
A season of statements in one pass
Host bookkeeping tends to happen once a year, in a hurry, shortly before a deadline — which is exactly when converting twelve PDFs one at a time is least appealing.
Batch processing takes up to 100 statements at once and merges them into a single sheet, with duplicate detection for overlapping periods and a source-file reference on every row, so a full year arrives as one sortable dataset.
Hosts running a separate account per property, or a personal account that quietly became a business one, get them consolidated into the same table rather than reconciled by eye across browser tabs.
When a lender asks what the property earns
Sooner or later a host needs to prove income — remortgaging, buying the next property, or satisfying a lender who is unconvinced that short-let income is real income. Lenders are notoriously sceptical here, precisely because the income is seasonal and platform-mediated.
What wins that argument is not a screenshot of a booking calendar. It is bank statements showing consistent payouts over time, converted into a clean monthly series that demonstrates the pattern rather than asserting it.
The same data answers the follow-up question, which is always about costs. A host who can show net earnings per property, evidenced against the statements, is having a very different conversation from one who can show a busy calendar.
What your accountant actually wants
Accountants who handle short-let clients ask for the same thing every year: the bank statements, as data, categorised, with nothing missing. What they usually receive is a folder of PDFs and a spreadsheet typed at the last minute.
Handing over a converted, balance-checked, categorised export changes the job. Your accountant spends their time on the judgement calls — the mortgage interest treatment, what counts as a repair versus an improvement, whether the furniture is capital — instead of on data entry billed at professional rates.
It also shortens the year end, because the query list that normally arrives in your inbox has largely been answered by the data before it was sent.
Straight into QuickBooks or Xero
Where you run accounting software, the converted statement can go straight in. FlowParse produces real bank-feed files — QBO, QFX and OFX — with a transaction ID per row, so a re-import does not double-post, and a Xero-ready CSV with the columns Xero expects.
That matters most for the accounts a live feed does not reach: the personal account the first property ran through, a second bank, the months before you set the business up properly. Those are precisely the gaps that otherwise get keyed in by hand.
Scanned and posted statements too
Older accounts and smaller banks still post paper, and the year you most need — the one before you started downloading PDFs — often exists only as a scan.
OCR runs first on scanned and photographed statements, then the AI structures the recognised text and flags low-confidence figures for a quick check, so a posted statement becomes the same clean rows as a downloaded one.
Numbers you can file a return on
Around 98% field-level accuracy on standard layouts, with every low-confidence figure highlighted in an editable preview before anything exports.
More importantly, the statement proves itself. Opening balance plus every transaction should equal the closing balance the bank printed, and FlowParse checks that on every statement. A missing payout or a duplicated row is caught arithmetically rather than discovered months later when the numbers will not tie out. That is what validation means here: evidence, not a confident-looking table.
Your account, your data
A host's statement is a personal document. It names your mortgage lender, your income, and often the account you also buy groceries from. Uploads run over TLS, processing is EU-hosted, the original PDF is deleted immediately after processing, and documents are never used to train AI models.
Nothing is retained once your export is produced.
Who this is for
Hosts with one property who want an evening of bookkeeping rather than a weekend, hosts with several who genuinely do not know which one earns, co-hosts who need their splits to be defensible, and the accountants who serve short-let clients and would like the payouts to arrive as data rather than as a PDF.
If you cannot currently say what a property earned net of its turnover costs last season, converting the statements is the step that produces the answer.
Convert your hosting bank statements
Upload a statement and get clean, categorised rows — payouts separated from fees, turnover costs visible, profit per property finally computable.