Batch processing

The same pipeline over many files with bounded concurrency, live progress, cancellation, and a per-file result rather than one failure aborting the run. This is the headless batch API driving a UI, not a separate code path.

Try it

Drop several images here

Pipeline

Longest edge 1600 px

Concurrency 2 Each worker holds a decoded image, so this is a memory dial as much as a speed dial.

Format

Convert to grayscale

Stop on first error

The batch API

ProcessBatchAsync is on ImageProcessor, which in a browser comes from ImageProcessorFactory.GetAsync() once the codec probe has run; on a server, AddBitBmageHeadlessServices registers it directly. Every item comes back with its own outcome - Output, Error or WasCancelled - and StopOnFirstError is opt-in.

var processor = await processorFactory.GetAsync();var result = await processor.ProcessBatchAsync(items, pipeline, new BatchOptions{    MaxConcurrency   = 2,    StopOnFirstError = false,        // one bad file must not lose the rest    ExportOptions    = exportOptions,    Progress = new Progress<BatchProgress>(p =>    {        completed = p.Completed;        if (p.Last.Error is { } error) Log(p.Last.FileName, error);    }),}, cancellationToken);foreach (var item in result.Succeeded) Save(item.Output!);foreach (var item in result.Failed)    Report(item.FileName, item.Error!);

Size MaxConcurrency against memory, not cores: four workers on 24 megapixel photos is roughly 400 MB of live pixels. Two is a safe default on phones.