Post-Render Transformers — Extensible HTML Processing Pipeline

The transformer pipeline lets you modify rendered HTML before it reaches consumers (email channels, PDF generators). This is particularly useful for email templates, where you need table-based layout, inline CSS, and Outlook conditional comments — concerns that template authors should not deal with manually.

IRenderedContentTransformer interface

public interface IRenderedContentTransformer
{
    int Order { get; }
    bool CanTransform(DocumentFormat format);
    Task<string> TransformAsync(string content, DocumentFormat format, CancellationToken ct);
}

Transformers run as an ordered mutation chain: each receives the output of the previous one and returns modified content. Use CanTransform to restrict a transformer to specific formats (e.g., HTML-only, skip PDF/Excel).

Registering a custom transformer

services.AddRenderedContentTransformer<HtmlMinificationTransformer>();

Writing a custom transformer

public sealed class HtmlMinificationTransformer : IRenderedContentTransformer
{
    public int Order => 300;
    public bool CanTransform(DocumentFormat format) => format == DocumentFormat.Html;

    public Task<string> TransformAsync(
        string content, DocumentFormat format, CancellationToken ct)
    {
        // Your minification logic here
        return Task.FromResult(Minify(content));
    }
}

Ordering conventions

Range	Purpose	Example
100	Structural transformation	MJML compilation
200	CSS processing	Inlining, purging
300+	Cosmetic	Minification, cleanup

The built-in MjmlTransformer (Order=100) runs first to convert MJML markup into table-based HTML. Application-level transformers should use Order 200+ to process the already-compiled HTML. See MJML Email Templates for details on the MJML transformer, and Rendering Pipeline for how transformers fit into the five-stage pipeline.