Trigger 
Describe Features
A page-by-page pipeline, end to end.
A page-by-page pipeline from PDF content streams to self-contained, machine-readable HTML.
Nothing in this path rasterizes the text layer. The optional native engine substitutes only the page background, and live HTML text still sits on top of it.
Text
Characters are grouped into runs by font/size/spacing, then optionally into
justified lines and paragraphs, and emitted as real <span>/semantic
HTML — never a screenshot.
Fonts
Each embedded font's glyph outlines are captured and rebuilt into a subset TrueType
font, then wrapped as WOFF 1.0 and embedded as a @font-face — so
rendered glyphs match the source document exactly, with no system-font substitution.
Background
Vector paths and raster images are emitted as an inline <svg>
mirroring the PDF's own graphics by default. For pages whose compositing an SVG
can't reproduce, an optional raster background renders the page graphics via the
native renderer while text stays live HTML on top.
Engines
The managed engine is fully managed, clip-aware, and needs no native library — it's the default. The optional native engine adds raster compositing and is available on 12 runtime identifiers.
Output shapes
Single self-contained file, one file per page, or a multi-page viewer with
navigation — all inline fonts/images/SVG as data: URIs, so there are no
outbound references and the artifact works offline.
Safety
Every dynamic string is HTML-encoded before it enters the output; link targets are
filtered to an http(s)/mailto/tel allow-list; hostile schemes
(javascript:, data:, file:) never reach the
page. Details on Security.
Choosing an engine
Managed is the default. Native is opt-in.
Both engines produce live HTML text. They differ only in how the page background is produced, and in what you have to deploy alongside your application.
| Managed Default | Native Optional | |
|---|---|---|
| Native library to deploy | None — pure .NET | Required alongside the application |
| Page background | Inline <svg> mirroring the PDF's graphics |
Raster compositing, text still live on top |
| Clip-aware text | Yes | Yes |
| Availability | Anywhere .NET runs | 12 runtime identifiers |
Review, not re-key
Because the output is real HTML, reviewers search, copy and comment on the converted filing directly — no round trip through the source PDF.