English原文：本指南暂无您所选语言的译本。
Short answer: Upload a PNG, JPEG or PDF drawing. For a PDF, pick the page with the parts table. Crop around the table, including its column headings, and select "Extract BOM." Then check every row against the drawing before you export to Excel or CSV. The tool reads an existing parts table with OCR, so the output is a draft that you verify.
Steps (per the tool's guide page)
1. Prepare the file. PNG, JPEG or PDF, up to 25 MB. PDFs must be unencrypted (up to 500 pages, one page and one crop read at a time). Images can be up to 60 megapixels. DWG, DXF and native SOLIDWORKS files aren't read directly, so export the sheet to PDF or PNG first.
2. Upload it, or select "Try a sample" to see the workflow first.
3. Pick the page that carries the parts table, for a PDF.
4. Crop the table with its column headings. Leave dimensions, notes and the title block outside the crop.
5. Select "Extract BOM." You can watch progress and cancel.
6. Review every row. Check part numbers character by character, then quantities and materials. Rows with a low OCR confidence label are flagged, but check unflagged rows too.
7. Correct cells directly in the editable table, which has the columns Item, Part number, Description, Quantity and Material.
8. Export to Excel (.xlsx, up to 500 rows) or CSV.
Worked example (generic)
An engineer has a scanned legacy assembly drawing with a 30-line parts list. They upload the PDF, pick page 1, crop the parts list with its header row, and extract. During review they fix two part numbers where the letter O was read as the digit 0, plus one quantity. They export CSV for the next step: comparing it with the current BOM.
Common mistakes
- Cropping without the header row.
- Including the title block or notes in the crop.
- Exporting without row-by-row checking.
- Expecting it to find components that aren't in the parts table. It doesn't infer a BOM from geometry.
Getting a clean crop
- One table per crop. If a sheet has a parts list split into two columns of rows, crop and extract each block separately, then join the exports in a spreadsheet.
- Multi-page parts lists. A PDF is read one page at a time, so extract each page's table and append the rows in order.
- Continuation headers. If the table continues on a second sheet without headings, include a heading row in your crop or rename columns after export so the files line up.
- Rotated tables. Some drawings place the parts list sideways. Rotate the page or image before upload so the text reads left to right.
- Low-quality scans. Rescanning at a higher resolution, within the 60-megapixel image limit, usually helps more than any later correction.
Edge cases to check during review
1. Lookalike characters. O and 0, I and 1, S and 5, and B and 8 are common OCR confusions in part numbers. Part numbers matter most, because the next tool matches on them exactly.
2. Merged cells. A description that wraps onto two lines can be read as two rows. Join them back into one row.
3. Quantity text. Entries such as "AR" (as required) or "REF" aren't numbers. Decide how you'll represent them before you compare, because BOM Compare needs plain non-negative numbers.
4. Columns your table has but the tool doesn't. Extra columns, such as vendor or finish, aren't in the editable table's columns. Add them after export if you need them.
5. Revision of the source. Write the drawing number and revision in the file name of the export, so the export can't be confused with another revision.
Turning the export into a usable BOM
- Keep the export as a record of what the drawing said, and make corrections in a copy if you need to normalize part numbers.
- If the part numbers will be matched with another BOM, normalize case and spacing consistently. BOM Compare matches exactly and is case-sensitive.
- Check for duplicate part numbers. A drawing table may list the same part twice in different positions. BOM Compare rejects duplicates, so sum them into one row or keep them in a format that tells them apart, depending on what you need to compare.
- The tool doesn't read DWG, DXF or native SOLIDWORKS files, so for a CAD-native BOM, export the BOM from the CAD system instead.
A second worked example (generic)
A supplier sends a two-page PDF assembly drawing. The parts list starts on page 1 and continues on page 2 without headings. The engineer extracts page 1 with its headings, then extracts page 2 with a crop that contains only rows. After export, they paste the page 2 rows under page 1 in a spreadsheet and check that item numbers run on continuously. Two rows on page 2 say "AR" in quantity. The engineer notes them for the purchasing team, and for comparison sets a placeholder agreed with the team, labeled clearly in the description.
FAQ
Is it free?
Yes. It is one of MST's free engineering tools.
Are my files used for training?
The tool page states that uploaded files aren't used for model training, and you can delete the server files from the tool.
What next?
To compare two revisions, export both as CSV and use the free BOM Compare tool.
Can it read a handwritten parts list?
The tool reads printed tables with OCR. Expect more corrections on handwritten or degraded tables, and check every row.
How many rows can I export?
Excel export supports up to 500 rows per the guide. CSV is the other option.
Does it read a BOM from a 3D model?
No. It reads a parts table on a drawing image or PDF.