How to turn document figures into a narrated whiteboard video

Short answer

Upload a figure-heavy PDF to docwhispr, pick the page range with the diagrams, and generate a narrated whiteboard explainer — without recording yourself. Edit slides and regenerate instead of reshooting a Loom.

The problem

Many technical documents are picture-first. Architecture diagrams, annotated schematics, training figures, and callout-heavy slides carry the meaning. Reading the surrounding paragraphs alone often is not enough — and neither is a 40-minute screen recording where someone scrolls, zooms, and rambles through a dense PDF.

What you usually want is a short, narrated lesson that stays focused on the figures: what to look at, in what order, and why it matters.

The method

With docwhispr, the workflow is intentionally narrow:

1. Upload a PDF that contains embedded figures (manuals, datasheets, and internal training docs work best).
2. Pick a page range where the diagrams live — you do not need the whole book.
3. Let the pipeline plan a lesson around those figures: extract images, write narration, and build whiteboard-style slides with highlights.
4. Watch the MP4, then edit slides in Excalidraw and regenerate if a callout or sentence needs tightening.

The product is picture-explanation-centric by design. Text still matters for context, but the lesson is organized around what is drawn on the page — not a wall of text-to-speech over every paragraph.

What you get

• A narrated whiteboard-style explainer video (MP4) you can share or archive.
• Slide images you can download as PDF or ZIP for handouts.
• Editable slides when narration or highlights need a second pass.

See a finished example: public sample project.

Step-by-step checklist

1. Choose the right source pages

Prefer pages where figures are central: flowcharts, block diagrams, timing charts, annotated screenshots. Skip pure legal boilerplate or blank front matter. Free plans are limited in pages and segments per job — start with the densest figure pages first. Details live in pricing and limits.

2. Generate, then skim for intent

After generation, ask: Does each slide highlight the right region? Does the narration match how you would explain the diagram to a colleague? If not, edit the slide and regenerate that part of the video instead of re-recording an entire Loom from scratch.

3. Share the outcome, not the raw PDF path

For onboarding and async review, the video plus slide pack is usually easier to consume than “open the manual at page 237.” Keep the PDF as the source of truth; use the explainer as the guided path through it.

When this is the wrong tool

If you need a live face cam, a mouse cursor touring a SaaS UI, or an interactive whiteboard meeting, a screen recorder is still better. See the comparison: docwhispr vs screen recording.

Related reading

• Explain datasheet and manual diagrams
• EDA / DFT manual figures → lesson clips