• A text-based PDF — one exported from CAD or Revit, where you can select the text in a PDF viewer. This is the one that matters: VOLTA reads the drawing text.
• Scans and outline-text PDFs are read by OCR on your device — slower, and tags can be misread, so check the takeoff. If a set cannot be read reliably, VOLTA says so instead of guessing: use the contact form and attach the PDF.
• Electrical sheets (the E-series) — and keep the legend, the panel schedules and the drawing notes in the set. Those pages carry most of what gets detected.
• The whole set at once — one combined PDF, one PDF per sheet (select them all together), or the ZIP as it came from the plan room. A sheet saved as a JPG or PNG is read by OCR.
• Text drawn as outlines. Some CAD exports draw the consultant's text as line-work (AutoCAD SHX fonts) — it looks like text but cannot be selected. AutoCAD keeps that text inside the PDF as hidden comments, and VOLTA reads it from there, exactly and with no OCR. If the comments were stripped (a flattened or re-saved PDF), those sheets are read by OCR instead.
• What it looks for: suite and unit counts, storeys, building count, panel designations, service amperage and transformer callouts on a residential set; rooms, workstations, meeting rooms, washrooms and mechanical tags on a commercial fit-out; luminaire tags on a retrofit — then it drafts quantities from those. Where the plans carry the fixture type tags as text, luminaires are counted against the luminaire schedule instead of estimated.

If a scan is all you have, you can still use VOLTA — open the drawings in the viewer and enter quantities yourself; the pricing, factors and quote all work the same.