Every page is read into its parts
Text, tables, charts, equations, drawings and code, each kept as its own part with the page it came from. A chart keeps its data, an equation stays maths, a drawing keeps each shape and label. Scanned and handwritten pages are transcribed; web pages are read as documents.
Your team decides: Which parts to use. The whole document, unless you pick sections, tables or drawings.
One real page, read. Pick a part to see what it became.
logoheadingheadingtext_blocktext_blockdrawingheadingtext_blockchartheadingtablecaptioncaption
Source: a quarterly operations review, 2 pages, as uploaded.
32 objects found · 1 chart · 1 table · 2 equations · 1 code block · 2 drawings · 1 image · 1 list · 8 headings · 9 text blocks · 5 captions · 1 logo
Read back as its five values in $M, not a picture of bars.
{
"type": "chart",
"values": [17.8, 19.4, 20.7, 21.5, 24.1],
"page": 1
}A draft is written from those parts, and only those
For a video, a storyboard: scenes with narration, each built from the parts it explains and pointing to their page. For a deck or a handout, the same parts laid out as slides or pages. Numbers, labels and formulas are taken from the page, never retyped.
Your team decides: What changes. Ask for it in plain words, scene by scene.
Your team reviews it before anything is made
The draft starts from good defaults. Change any of them, and approve it on the same board as your teammates.
Your team decides: The look, the voice, the presenter, the brand, and when it is ready.
Shape the storyboard by talking
Ask for changes scene by scene. Teammates review and approve on the same board.
Starts from a draft built from the document's own parts.
Choose the look
Templates made for each use case, or your own PowerPoint template.
Uses your last choice, or one made for the use case.
Choose the voice, presenter and music
Narrators from Azure, ElevenLabs, Sarvam or OpenAI, a presenter on screen if you want one, and music at the level you set.
Starts with an Azure narrator. Only promos start with music.
Use your brand kit
Your logos, colours and fonts, read from a document you already have.
Uses your workspace's brand kit.
The files are made at the price you saw
Video from 480p to 1080p, with PowerPoint and PDF from the same draft. Every run shows its price before it starts. Change one scene later and only that scene is made again.
Your team decides: Quality and files. Nothing runs until you start it.
The same parts answer, compare and connect
The document is read once. Everything else uses that reading.
Your team decides: Who can ask, and which documents an answer may come from.
Ask
Answers that quote the exact page, table or chart they came from. If the document doesn't say, the answer says so.
Compare
Two versions of a policy, SOP or manual, part by part: what was added, removed or changed.
Connect
Your AI agents get the exact clause or table they need through the API or MCP, not the whole file.
The API and MCP use the same prices as the app. See pricing
When the document changes
Upload the new version. Truframe measures what changed between the two files and shows it part by part. Every video and deck made from the old version says so, with the pages that changed. Make it again from the new version in one step; the old one is kept.
Your team decides: What to make again, and when.
Bring one document. See what it becomes.
Send us a chapter, SOP or policy you already use. We set up a workspace with trial credits for it. After that, pay only for what you make.
Prices for every output are on the pricing page.