ITPro Address Capture · Local PDF to Excel
Automatic extraction of addresses and other data from documents.
Read addresses and other fields from PDF into Excel. Create document profiles and check results against the source.
Create and edit your own profiles. Our help with configuration and testing is optional.
Inside your environment
Your local processing boundary
- 01Your PDF documentsDigital documents or scans saved as PDF.
- 02Local text reading and OCRThe program chooses the reading method automatically.
- 03Optional local AI analysisAddress suggestions checked by your team.
- 04Your Excel workbookPrepared and stored in your environment.
Document processing stays inside this boundary.
What changes for your team
PDF to Excel, with profiles you control.
Profiles your team controls
Create and edit document profiles in the built-in visual editor. Select the areas to read, define your fields and reuse saved profiles for recurring work. Our help with configuration and testing is optional.
Keep document data in your environment
PDF reading, OCR and optional AI assistance with assigning address text to fields take place locally. Documents and extracted information do not need to be uploaded to an online converter or cloud AI service for processing.
Less retyping, clearer checks
For exports that include address images, the Excel workbook brings the extracted address fields and cropped images of the scanned address areas together. Compare the values with their source and correct them directly in Excel, without repeatedly reopening the PDFs.
An illustrative time comparison
100 address records. Much less manual work.
Compare around one hour of manual entry with recurring processing using a prepared profile. This example assumes 100 documents with one address record in each.
Manual entry
~60 min
Baseline: one person enters around 100 address records per hour.
Existing PDF files
~3 min
Around 57 minutes saved per 100 records — approximately 95% less time.
Scanning and processing
~6 min
Includes physical scanning, program processing and routine wrap-up: around 54 minutes saved — approximately 90% less time.
Even greater potential when staff must search through document text. Finding each required field and then retyping it adds manual work. A prepared profile reads the selected areas repeatedly, reducing both the search and the copying effort.
Indicative timings for recurring work with a prepared profile. Actual times depend on document quality, scan handling and the checks or corrections required. Additional fields and more complex layouts should be assessed with representative documents.
The business workflow · Processing stays local
From PDF to Excel, inside your environment
Create and reuse profiles for your document layouts, or ask us to help with setup. The program chooses text reading or OCR automatically, with optional local AI support for assigning address text to fields. Your team stays in control of the result.
01
Choose your PDF
Open a digital, scanned or mixed PDF inside your environment. One file can contain many letters, recipients or document sets.
02
Create or select a profile
Mark the areas to read, choose your fields and save a profile for that document layout. Reuse it next time. Configure it yourself or ask us to help.
03
Automatic text or OCR
The program first checks for usable PDF text in the selected areas. If text is missing or unreadable, it switches to OCR locally.
04
Review address fields
Address rules provide the initial split. For selected uncertain results, optional local AI can propose a clearer field structure. Your team can keep, accept or edit the result before export.
05
Review and use your Excel workbook
Create an XLSX workbook locally with the columns defined in your profile. Where the export includes address images, compare the source beside the extracted data and correct values directly in Excel. Your team reviews the results and controls any onward sharing.
Prepare and test a profile once, then reuse it for recurring documents. Configuration support is available if you need it. PDF content, extracted fields and any information used by the optional local AI stay inside your environment during processing.
What local AI actually does
Reading the words is only half the job.
A recipient’s name may wrap onto two lines, followed by a street and number with no “ul.” or “al.” prefix. Copying each line into the next Excel column can put part of a name into the street field — even when every word was read correctly.
The challenge is to identify what each part means. Digits in a name, missing characters or lines merged during reading can make the boundary between recipient and address unclear. The program starts with address rules; optional local AI helps interpret selected uncertain results in context.
Text read from the document
Ośrodek Przykładowy
Sekcja Dokumentacji
Kwiatowa 12/4
00-123 Warszawa
The first two lines form one recipient name. The final line contains two fields: postal code and city. The task requires grouping and splitting, rather than copying lines into columns.
The field structure to check
RecipientOśrodek Przykładowy Sekcja Dokumentacji
Street and building / flat numberKwiatowa 12/4
Postal code00-123
CityWarszawa
Fictional Polish address illustrating the expected field grouping. Many such layouts can be handled by address rules alone; local AI assists with selected uncertain results.
Rules first. Local assistance for the uncertain cases.
Not every record goes to AI. When local AI review is enabled, the program selects results using the configured checks and asks the operator whether to run it. The model receives the extracted address text, initial fields and review reasons — not the whole PDF.
01
Read the address block
Read the selected PDF area as text, or use local OCR when needed.
02
Apply address rules
Group lines into recipient, street with number, postal code and city.
03
Select uncertain results
Configured checks flag issues such as an unclear street or low reading confidence.
04
Ask local AI, if enabled
The operator can start local analysis of the extracted block and the initial fields.
05
Review, then export
Keep the initial result, apply a proposal or edit the fields before producing Excel.
Local AI proposes. Your team decides. The model can suggest which extracted words belong to the recipient, street with building or flat number, postal code and city. Your team compares the extracted text, the initial fields and the proposal, then keeps, accepts or edits the result.
Review is configurable: the program does not force a separate approval click for every record. Suggestions are not a postal-register check and still need appropriate source checks before use.
Address analysis stays on your computer. The local model does not require the address text to be uploaded to a cloud AI service.
What can be collected
Extract any text fields your process needs.
You choose the information to capture by marking the areas to read on your PDF. Use the built-in editor to define those areas as fields, choose reading rules and Excel columns, then save your own reusable profile. The examples below are starting points — your fields are defined by your documents.
Recipients and addresses
Names, street and building or apartment number, postal code and city — from the address areas you select.
Company details and tax IDs
Company names, tax ID (NIP) and other business identifiers present in the selected document areas.
Document and case references
Document numbers, case identifiers, order references or other codes your team uses to identify a record.
Dates
The issue date, a deadline or another date shown in the area configured for that field.
Amounts and figures
Amounts, quantities or other text-based figures from the selected parts of a recurring document layout.
Your own text fields
Other text information your process needs. You choose the areas, field names and the corresponding Excel columns.
Test your selected areas on representative documents to check that the text can be read reliably. You can also define how a combined PDF is grouped into records, using a page count or document markers. We can help configure and test profiles if you wish.
Built for address processing, including the final check
The address image and extracted data. Together in Excel.
When address images are available in the selected export mode, the Excel workbook includes cropped images of the scanned address areas alongside the extracted fields. Your team can compare each result with its source and correct names, streets or postcodes directly in the workbook, without searching for the matching letter or switching back to the PDF.
Purpose-built for this task. ITPro Address Capture goes beyond recognising text: it turns address documents into a structured, editable list with the visual evidence needed for practical quality checks. Less switching between files, a clearer review process and more control over the data before it is used.
From selected areas to structured data
See the data and its source. Correct the result in Excel.
Compare the address image on the left with the extracted fields on the right. If a character was misread, correct the value directly in Excel. The source stays beside the data throughout the check.
Illustrative view with fictional data, showing columns D–H of the “Adresy” worksheet in their actual order. The export uses the Polish headings shown below; English explanations are added for clarity. Other columns are omitted from this focused example.
On smaller screens, each card contains one record. The source image appears first, followed by the extracted fields.
| RawOCR (crop)Address image + raw text | AdresatRecipient | Ulica / nrStreet and number | KodPostal code | MiastoCity |
|---|---|---|---|---|
 | DEMO COMPANY ALPHA | Example Street I2/4Compare with image | 00-000 | Demo City |
 | DEMO COMPANY BETA | Sample Avenue 8 | 00-000 | Demo Town |
One character. One quick correction.
The address image reads “12/4”. The initial extracted text contains “I2/4”. In Excel, replace the letter I with the number 1 in column F, “Ulica / nr”.
The button illustrates a manual edit of the exported workbook. It does not rerun address validation or change the original OCR text.
- 1. CompareRead the address image beside its fields.
- 2. Correct in ExcelEdit the misread value in the workbook.
- 3. Use the checked dataContinue with the corrected address list.
Shown here: an export with an embedded address image available. Other export columns contain record, page, profile and review information. A document profile can add fields such as a tax ID (NIP); NIP is not a standard column.
Where it can help
Practical uses for recurring document work.
Correspondence and mailing lists
Build a recipient list from recurring letters in a combined PDF. Review names and postal addresses before using the list in your correspondence process.
Case and document registers
Collect selected references, dates and text fields from repeat document layouts. Prepare a consistent register without retyping the same information.
Recurring order or invoice layouts
Extract selected header fields from a known layout. Full line-item reconstruction or automatic accounting entries need a separately assessed workflow.
Fit and expectations
Plan the result before you start.
The best starting point is a recurring document layout and a clearly defined result. Handwriting, arbitrary tables and every possible PDF format are not assumed to work automatically.
For each document workflow, consider:
- Representative PDFs, including difficult examples.
- Which fields matter and how pages form a record.
- The required workbook layout and review rules.
- Volumes, frequency and any downstream systems.
Local processing and data control
Keep document processing inside your environment.
Document content, extracted fields and information used for optional AI-assisted address field analysis are processed locally. This reduces the need to expose sensitive documents by uploading them to external online converters or cloud AI services.
The optional local AI assistant helps interpret uncertain address text, such as wrapped recipient names or unclear boundaries between a name and a street. It proposes field assignments; your team can compare them with the extracted text, retain the initial result or make corrections.
Licensing and updates use a separate network connection; PDF content and extracted fields are not sent to the licensing service. Optional reporting, support sharing or onward integrations are configured separately and agreed with you.
Configure it yourself, or ask us to help. Your team can create, test and update its own document profiles. We can optionally help with initial configuration, testing and training, or scope additional integrations to fit your business process.
Environment and licensing
Environment: Windows 10/11, 64-bit (x64). Licensing is based on actual usage. Commercial terms, installation details and hardware requirements are available on request.
Document contents and extracted data are processed locally. Activation and transmission of usage metadata require a connection to the licensing service.
Common questions
Questions about ITPro Address Capture
Can we create and change document profiles ourselves?
Yes. The built-in visual editor and setup wizard let your team mark the areas to read, define fields, set page grouping and choose the Excel layout. Save profiles for reuse, add new layouts and adapt existing profiles as your requirements change.
Can ITProSupport help us with configuration?
Yes, if you would like assistance. We can prepare and test profiles for your documents, help your team get started and assess any additional integration needs. This is optional support; your team can also configure and manage its own profiles.
Does the program choose between PDF text and OCR automatically?
Yes, in Auto mode. It first checks whether the selected PDF areas provide usable text. If the text is missing or unreadable, it uses OCR to read those areas locally. Different areas of a mixed PDF can use different reading methods.
What does local AI actually do with an address?
Address rules provide the initial fields. For selected uncertain results, optional local AI uses the extracted address text, those fields and review reasons to propose how words belong to the recipient, street with building or flat number, postal code and city. This can help with wrapped names, unclear line boundaries and streets without identifying prefixes.
Your team can retain the initial split, apply a proposed change or edit the fields before export. Separate approval of every record is not mandatory. Analysis stays local; it is not a postal-register check or a guarantee of correctness.
Can it read scanned documents?
Yes, when the scan is supplied as a PDF. OCR reads the selected areas locally. The quality of the scan and document layout must be checked on representative samples. Standalone image files need conversion to PDF first.
Does every document work without setup?
Recurring document layouts use saved profiles. Your team can create and test them on representative PDFs, with optional help from us. A new layout, page grouping or business rule may require a new or updated profile and another test.
Can the result go to CSV, CRM or accounting software?
The native result is an Excel XLSX workbook. A CSV copy can be prepared from the workbook. Connecting the result to other systems is a separately scoped service.
Can the whole folder run automatically?
The application processes one PDF per run and can handle many document records within that file. Folder monitoring, email intake and recurring multi-file processing require additional workflow setup.
Will every value be correct automatically?
No text-recognition workflow should be treated as an unchecked source of truth. Scan quality, fonts and layout affect the result. Your team decides which fields need review and tests each profile against sample documents. We can assist if needed.
Explore ITPro Address Capture with your own documents.
We supply a configurable tool that keeps document processing in your environment. Create your own profiles for the documents and fields you need. If you prefer assistance, we can help with configuration, testing and getting your team started.