Branching Pipelines

The Recast pipeline is immutable — every method returns a new pipeline instance without modifying the original. This means you can build a shared base and fork it into multiple outputs.

Basic branching

import { Recast, OpenAIProvider } from 'playwright-recast'

// Shared base: parse + speed processing
const base = Recast
  .from('./test-results/trace.zip')
  .parse()
  .speedUp({ duringIdle: 3.0, duringUserAction: 1.0 })

// Branch A: English voiceover
await base
  .subtitlesFromSrt('./en.srt')
  .voiceover(OpenAIProvider({ voice: 'nova' }))
  .render({ format: 'mp4' })
  .toFile('demo-en.mp4')

// Branch B: Czech subtitles only (no voiceover)
await base
  .subtitlesFromSrt('./cs.srt')
  .render({ burnSubtitles: true })
  .toFile('demo-cs.mp4')

Both branches share the same trace parsing and speed configuration. The base pipeline is not modified — each branch creates its own chain.

Multi-language voiceover

Generate localized versions from the same trace:

import { Recast, OpenAIProvider, ElevenLabsProvider } from 'playwright-recast'

const base = Recast
  .from('./test-results/trace.zip')
  .parse()
  .speedUp({ duringIdle: 3.0 })
  .autoZoom({ inputLevel: 1.4 })
  .clickEffect({ sound: true })

// English — OpenAI
await base
  .subtitlesFromSrt('./srt/en.srt')
  .voiceover(OpenAIProvider({ voice: 'nova', speed: 1.1 }))
  .render({ format: 'mp4', burnSubtitles: true })
  .toFile('demo-en.mp4')

// German — ElevenLabs
await base
  .subtitlesFromSrt('./srt/de.srt')
  .voiceover(ElevenLabsProvider({
    voiceId: 'onwK4e9ZLuTAKqWW03F9',
    languageCode: 'de',
  }))
  .render({ format: 'mp4', burnSubtitles: true })
  .toFile('demo-de.mp4')

// Japanese — subtitles only, no voiceover
await base
  .subtitlesFromSrt('./srt/ja.srt')
  .render({ burnSubtitles: true })
  .toFile('demo-ja.mp4')

Resolution variants

Generate multiple resolution outputs from one pipeline:

const base = Recast
  .from('./test-results/trace.zip')
  .parse()
  .speedUp({ duringIdle: 3.0 })
  .subtitlesFromSrt('./narration.srt')
  .voiceover(OpenAIProvider({ voice: 'nova' }))

// 1080p — standard
await base
  .render({ format: 'mp4', resolution: '1080p', burnSubtitles: true })
  .toFile('demo-1080p.mp4')

// 720p — lighter weight
await base
  .render({ format: 'mp4', resolution: '720p', burnSubtitles: true })
  .toFile('demo-720p.mp4')

// 4K — high quality
await base
  .render({ format: 'mp4', resolution: '4k', burnSubtitles: true })
  .toFile('demo-4k.mp4')

With and without branding

Generate branded and clean versions:

const base = Recast
  .from('./test-results/trace.zip')
  .parse()
  .speedUp({ duringIdle: 3.0 })
  .subtitlesFromSrt('./narration.srt')
  .voiceover(OpenAIProvider({ voice: 'nova' }))

// Branded version — intro, outro, background music
await base
  .intro({ path: './assets/intro.mp4', fadeDuration: 800 })
  .outro({ path: './assets/outro.mp4', fadeDuration: 800 })
  .backgroundMusic({ path: './assets/music.mp3', volume: 0.25 })
  .render({ format: 'mp4', burnSubtitles: true })
  .toFile('demo-branded.mp4')

// Clean version — no branding, no music
await base
  .render({ format: 'mp4', burnSubtitles: true })
  .toFile('demo-clean.mp4')

Parallel execution

Since each branch is independent, you can run them in parallel:

const base = Recast
  .from('./test-results/trace.zip')
  .parse()
  .speedUp({ duringIdle: 3.0 })

await Promise.all([
  base
    .subtitlesFromSrt('./en.srt')
    .voiceover(OpenAIProvider({ voice: 'nova' }))
    .render({ format: 'mp4' })
    .toFile('demo-en.mp4'),

  base
    .subtitlesFromSrt('./cs.srt')
    .render({ burnSubtitles: true })
    .toFile('demo-cs.mp4'),
])

Note that each branch re-executes the full pipeline from source. The base object is a pipeline description, not cached intermediate results. For large traces, consider whether parallel execution is worth the CPU cost.

Next steps

Minimal Pipeline — start simple before branching
Full Pipeline — see all stages in one pipeline
Recast Class API — immutability and method reference