Use case

Cash flow forecasting for the weeks that actually matter

Forecast weekly net cash flow from a CSV, use the band to see the downside, and plan payroll and supplier payments with a visible range.

Run the sample Put this in production

A cash flow forecast is judged by its downside, not its median. The question is not “what will the balance be?” but “how low could it go before the next receivable lands?” A range makes that question explicit.

The sample is weekly net cash flow, including a payroll cycle. Negative values are allowed; the model does not require a positive series. If you forecast the bank balance instead of the flow, the trend will dominate and the weekly cycle will be harder to see.

The sample series

The studio below is preloaded with this weekly net cash flow (usd) series and a 8-step horizon. It is an illustrative example, not a real company’s data. Replace it with your own CSV to run the same workflow.

Show the CSV

week,net_cash
2025-11-03,18200
2025-11-10,-12400
2025-11-17,9600
2025-11-24,-15800
2025-12-01,22400
2025-12-08,-11900
2025-12-15,13100
2025-12-22,-17600
2025-12-29,8400
2026-01-05,-14200
2026-01-12,19600
2026-01-19,-12800
2026-01-26,11200
2026-02-02,-16400
2026-02-09,24100
2026-02-16,-12100
2026-02-23,14900
2026-03-02,-17100
2026-03-09,10300
2026-03-16,-13800

CSV forecast studio

Forecast your cash flow forecasting series.

This studio is preloaded with the weekly net cash flow (usd) example above. Replace it with a CSV of your own history, choose a horizon, and run the forecast. The public demo is free while the Hugging Face Space is available; the API is the paid product.

1 · Paste or upload a CSV2 · Pick a horizon3 · Download the 21 quantiles

Steps

1 credit per API call · the web demo is free while the public Space is available

What to watch in the output

Forecast flows, not balances

A balance series is cumulative and dominated by its starting level. A flow series shows the weekly pattern that actually creates the risk.

Put payroll and tax dates in the calendar

The model learns the pattern in the history. If a new payroll cycle starts, the old pattern is no longer the whole story; model it as a scenario.

Use the low band for the liquidity buffer

The low band is the planning downside. Compare it with the minimum cash policy, not with the median.

Common mistakes

—Mixing bank balances and flows in one series.
—Ignoring a large one-off payment. It will look like a new level and distort the next few weeks.
—Forecasting too far ahead without a receivables model. The band will widen, which is honest but not a substitute for collections work.

How to put it to work

1. Start with the sample. Run the preloaded series once so you see the output shape before you debug your own CSV.
2. Replace it with your history. Keep one timestamp column and one or more numeric columns. Mark outages and stockouts as empty cells, not zeros.
3. Backtest before you plan. Hold out the last few periods, compare the forecast with what happened, and check whether the actuals land inside the 80% band.
4. Move it into production. Use the REST API for a nightly job or an internal tool. One credit per successful call; failed calls are refunded.

FAQ

Can the model handle negative values?

Yes. Negative cash flow is a valid value. The “keep forecasts at or above zero” option should stay off for net cash flow.

Should I use weekly or monthly data?

Use the cadence of the decision. Weekly if payroll and supplier payments drive the risk; monthly if you only need a board-level view.

Can I include receivables as a second column?

You can include related numeric columns, but the model does not enforce an accounting identity. Treat the output as a statistical forecast, not a ledger.

Need this as an API?

The same forecast is available as a metered REST endpoint with JSON and a downloadable quantile CSV.

See the API