Skip to content
JARAI Help

First production walkthrough

This walkthrough takes you from a logged-in Studio Console to a finished, downloadable video. Plan for about 15–20 minutes of clock time (most of it waiting on the AI pipeline to run).

The walkthrough

  1. Open the New Production page.

    From the left sidebar, click Productions → New Production. You’ll land on a form titled “Create a new production”.

  2. Pick a content theme + project type.

    • Project type (e.g., “Short-form video”, “Podcast episode”, “Book chapter”) sets the pipeline shape: which Steps activate, what AI providers are called, what the final deliverable looks like.
    • Content theme (e.g., “Tech news”, “Recipes”, “Historical events”) narrows the brief generator’s subject space.

  3. Pick an avatar (or two).

    For solo productions, pick a single avatar — that voice + face will narrate. For dialogue formats, pick Character 1 + Character 2 avatars and the script generator will produce alternating-speaker scripts.

  4. (Optional) Override the subject.

    The brief generator picks a topic for you within the content theme. To force a specific topic, type it into Subject override (e.g., “How do solar panels work in winter?”). Otherwise leave blank and JARAI invents one.

  5. Click “Start production”.

    The production lands in your queue with Status = Queued. Within a minute, Step 1 (ProductionBriefGenerator) starts and you can watch progress on the Production detail page.

  6. Watch the pipeline run.

    The production detail page shows each pipeline step with a live status badge:

    • Pending (waiting for upstream)
    • InProgress (pulsing — running now)
    • Complete ✅
    • Failed ❌ (with retry button)
    • PendingReview (gated on operator decision)

    Steps 1–4 (text generation) take ~1–2 min total. Steps 5–6 (image acquisition + analysis) ~3–5 min. Steps 7–11 (image/voice/video production) ~5–10 min depending on length. Steps 12–15 (variants, attribution, gates) ~1–2 min.

  7. Review the result.

    When Step 14 (DeliverablePackager) completes, the brief enters Status = Awaiting publish. You can:

    • Preview every asset inline (thumbnails, audio players, video posters) on the Story tab.
    • Download individual assets or all assets as a zip.
    • Approve to publish (sends to Loop 2 distribution) or Reject (deletes the deliverables).

  8. (Optional) Approve to publish.

    If you chose target platforms in step 2, hitting Publish triggers Loop 2 (DistributionOrchestrator) which dispatches the platform-variant videos to LinkedIn, TikTok, etc. Each platform shows a publication status badge after dispatch.

What you’ll see at each stage

StageWhat’s happeningWhat you can do
Steps 1–4Brief → Narrative → Beats → Scenes (text generation)Read the generated story on the Story tab; cancel if it’s wrong
Steps 5–6Image acquisition (Pexels / Pixabay / Wikimedia) + analysisReview per-scene image candidates; override the chosen image if you want
Steps 7–10Per-scene image production + voiceover TTS + scene-video compositionPreview audio + video stills on the Story tab
Steps 11Platform variants (16:9 / 9:16 / 1:1) renderedPreview each variant inline
Steps 12–14SEO metadata + attribution + final packagingApprove to publish, or download for manual upload

Where to learn more