Payslip data extractionfrom any employer.
Gross pay, net pay, contributions and deductions from any employer's payslip layout, normalized into one schema. The same value has many names across issuers and languages; the extraction knows they are one field.
HR and lending teams read any employer's payslip as the same clean record.
What MiruIQ extracts from a payslip.
The typical payslip schema, normalized across issuers.
| Field | What it captures |
|---|---|
| Employee | Name and identifiers of the employee. |
| Employer | Issuing company, the layout owner. |
| Pay period | Month or period the slip covers, normalized. |
| Gross pay | One field regardless of wording: Bruttolohn, Lohn, salario, gross pay. |
| Net pay | The paid-out amount, cross-checked against the calculation. |
| Social contributions | Pension, unemployment and insurance contributions, itemized. |
| Tax withheld | Income tax deducted at source. |
| Allowances and bonuses | Recurring and one-off additions to base pay. |
| Deductions | Everything subtracted between gross and net. |
| Year-to-date totals | Cumulative figures where the slip carries them. |
| Payment IBAN | The account the salary is paid to. |
Not a pre-trained model, a schema you control.
Payslips are the document type where templates die fastest: thousands of employers, each with its own layout and vocabulary. Semantic schemas extract by meaning instead, so one definition covers them all.
The figures also have to reconcile: gross minus deductions must equal net. Recalculation catches both extraction errors and documents where the numbers genuinely do not add up, and those park for a human.
Payslip extraction questions
What HR, lending and verification teams ask most.
You define the payroll fields once (gross, net, contributions, period) and extraction finds them by meaning on any employer's layout. Figures that must add up are recalculated, and slips that fail a check park for human review. Normalized records deliver to your HR, payroll or lending systems.
No. There are no per-employer templates and no pre-trained payslip model to maintain: extraction is generic and schema-driven, so a payslip from an employer you have never seen processes like any other. Bruttolohn, Lohn and salario all land in the same gross-pay field by meaning.
Yes. Payroll data is among the most sensitive a company holds, and MiruIQ deploys accordingly: on-premise with local AI models, in Swiss hosting, or in the cloud, your choice per pipeline. Nothing about the workflow requires documents to reach a third-party AI provider.
See it on your own documents.
The 14-day free trial covers 200 documents: enough to define your schemas, run real files and measure the results before anyone calls you.