Lightweight HTML/CSS rendering engine written in pure Free Pascal

[dsiders]

In that case, qt6 only, I'm not willing to change bindings for Qt5 since it is EOL, but anyway, QSvgRenderer can be added into it's own unit, so won't kill current lazarus users so they must rebuild C bindings. In a few years when distros update bindings then such unit can be merged to the qt62.pas

QSvgRenderer would certainly be useful for Shell controls too.

[mattias]

Impressive!

I just implemented flex and grid layout for Fresnel, so I know how much work and trouble it is. Please make Pixie open source, so Pixie and Fresnel can learn from each other. The main difference is that Fresnel uses components as elements, editable in the Lazarus designer.

[Gustavo 'Gus' Carreno]

Hey Wallaby,

I'm considering open-sourcing it and wanted to see if there's interest in the community first.

Please, PLEASE, PLEASE open source this !!!

I've been looking forward for something better than TFrameView for so long that I kinda lost hope. Then YOU come along and present what looks like a very, VERY professional looking piece of software that is just bloody AWESOME!!!!!!!

Thank you so, so very much in case you do decide to open source it!!

Cheers,
Gus

[kirchfritz]

Please open source it!

[JD]

Would there be interest in seeing this released as open source?

+1

It looks impressive. Well done.

JD

[speter]

Would there be interest in seeing this released as open source?

Yes please. I think it is a safe bet that there would be lots of interest.

cheers
S.

[szlbz]

It would be even better if it could be open sourced

[egsuh]

I love this.

But doesn't HTML/CSS rendering engine convert HTML/CSS to Lazarus LFM files?

As many are saying THtmlViewer is not sufficient, I'd like to recommend TWebView4Delphi to simply display web pages within Lazarus application. The most important thing is that this requires only one file, webview2loader.dll, to distribute. It can display all recent web pages.

[Wallaby]

Thanks! To clarify — Pixie has nothing to do with LFM or converting HTML to Lazarus forms. It's a rendering engine: you give it HTML, it parses, lays out, and paints the result directly onto a canvas. The output is pixels on screen, not Lazarus components.

Regarding WebView2 — it's a solid option if you need a full browser, but it solves a different problem:

Cross-platform. WebView2 is Windows-only. Pixie runs identically on Windows, Linux, and macOS from a single codebase.

Zero dependencies. Even "just one DLL" is still a dependency — and WebView2 also requires the Evergreen Runtime on the target machine. Pixie compiles into your executable. Nothing to ship, nothing to version-check.

No out-of-process complexity. WebView2 runs Chromium in a separate process, bringing IPC overhead, a user data folder, async message passing, and a whole class of bugs outside your control — Edge updates, runtime mismatches, sandbox issues, enterprise GPOs blocking it, antivirus flagging the broker process. Pixie renders in-process, in your app's paint cycle.

Lightweight. An embedded browser pulls in a JS engine, networking stack, and media pipeline to display what's often just styled text and tables. Pixie does that at a fraction of the footprint.

Native integration. Pixie renders via Direct2D/Cairo/CoreGraphics, so you get proper DPI scaling and theme integration without the embedded-foreign-window issues (focus, z-order, transparency, resize flicker).

If you need to load arbitrary websites with JavaScript — use a WebView. If you want HTML/CSS as a lightweight way to render rich content in a native app, that's what Pixie is for.

[440bx]

If you want HTML/CSS as a lightweight way to render rich content in a native app, that's what Pixie is for.

Absolutely no doubt that for HTML/CSS only rendering it is a significantly superior solution than any other that implements a full browser.  Not to mention its "no dependencies" and portability.

Aside from that, so far, unsurprisingly, it seems unanimous that making it open source would be a very welcome move.

[Wallaby]

Cool, I'll publish it shortly! Just need to clean it up a little, add a demo and organise the files. Stay tuned :-)

[440bx]

Cool, I'll publish it shortly! Just need to clean it up a little, add a demo and organise the files. Stay tuned :-)

Sounds great!

I'm sure I'm not the only one who is going to be very pleased.  Great news, thank you!.

[egsuh]

Thanks! To clarify — Pixie has nothing to do with LFM or converting HTML to Lazarus forms. It's a rendering engine: you give it HTML, it parses, lays out, and paints the result directly onto a canvas. The output is pixels on screen, not Lazarus components.

I see. I misunderstood. This is a great tool anyway.

[440bx]

@Wallaby,

Just a suggestion... now that you've decided to make Pixie open source, I believe hosting the source code at the SoftPerfect site (instead of github or some other repo), would provide greater exposure to SoftPerfect's products and free utilities.

On a different note, your company has a number of interesting products.  I think the utility that redirects a browser's cache to a RAM disk is a particularly good idea.  Browsers are IMO the most performance damaging apps. They eventually use gigabytes of disk space in hundreds of directories, they never clean up after themselves.   Because of that, I wrote a utility to get rid of the browser's cache (chrome only, because that's what I use) and other temp files that are always left laying around.

[Wallaby]

Thanks for the suggestion! I think GitHub is the gold standard for open source though — it makes it much easier for everyone to submit issues, pull requests, and contribute. Hosting it elsewhere would be convenient for me but inconvenient for everyone else, and that's the wrong trade-off for an open-source project.

I'd rather use GitHub and just add a couple of links to my website, which you are of course welcome to browse and maybe even pick up some tools if you find them useful

And yes, the RAM Disk + browser cache combo works really well — browsers are notorious for thrashing the disk with gigabytes of temp files across hundreds of directories. Redirecting that to a RAM disk keeps things fast and clean without any manual maintenance.