How to Create LONG AI Videos in One Prompt (Invideo + VEO 3.1)

Summary of the Tutorial (Technological Concepts, Product Features, Workflow)

• Core claim/problem solved: Generating AI video has often been limited to very short clips (8–12 seconds), which forces users to stitch clips together—leading to problems like inconsistency and pacing issues. This tutorial presents a workflow designed to generate long, full-length AI videos (up to ~10 minutes) in a single generation.

• Main tool: Invideo (shown as “Nvido” in subtitles) is positioned as (per the speaker) the only AI video generator that can create long videos in one go.

• Model selection for consistency: The user selects a high-end video model (the highest available). The speaker emphasizes that for longer videos, consistency matters more.

Workflow automation (no manual prompting)

• The key feature is the “Workflows” tab, which provides pre-made workflows.
• Instead of writing prompts manually, the workflow handles the prompting internally, producing one complete prompt based on your settings.

Video types and storytelling templates

• Workflows are grouped into categories such as:
  • Generated films
  • Faceless videos
  • Marketing videos
  • UGC ads
• For long-form output, the speaker recommends Generated Films or Faceless Videos.
• The example uses Faceless Explainer → Script-to-Video, where you paste:
  • Topic/title
  • Description/outline
• This approach automates long-form structure, reducing time spent assembling a coherent sequence.

Configurable generation length and platform

• Length selector: up to 10 minutes (example set to 3 minutes).
• Platform selector: example given is YouTube, which tailors output for posting.

Style and engagement settings

• Background music: choose by describing a vibe (example: calm instrumental music) rather than selecting a specific track.
• Subtitles: options include bold subtitles with a popping effect to improve retention.
• Watermark: a toggle to turn watermark off.

Voiceover configuration

• Select a voice actor (example: clear American female narrator) for narration across the entire video.

Asset sourcing (“stalker generative” setting)

You can choose what assets the system uses:

• Only generated clips (fully AI-generated scenes)
• Only generated images
• Only stock media (the speaker mentions a Nvidia stock library)

For the demo, the speaker selects only generated clips to create an all-AI video.

Two-stage generation approach

1. Generate (initial processing) using the workflow-created prompt.
2. Generate media using a Pro setting to use “best models,” such as:
   • Google “VO 3.1”
   • Sora 2

Review/quality assessment (what worked)

The speaker highlights:

• Visual quality (implying high-end generative models)
• Music fit with the chosen vibe
• Clarity of information delivery
• High-end sounding voiceover

---

Editing Capabilities (How Revisions Work)

The tutorial explains two entry points:

• Editing mode: for larger/broader changes.
• Gallery mode: shows scenes in order.

Prompt field edits

• You can adjust settings such as aspect ratio and even translate the entire video.
• Example: changing language to French and triggering regeneration.

Edit toggle with four categories

1. Media
   • Edit scenes individually
   • Replace with:
     • uploaded media
     • stock media
     • new AI-generated clips
2. Music
   • Upload a track
   • Switch to stock music, etc.
3. Script
   • Edit line-by-line
   • Changing the script can regenerate matching media
4. Settings
   • Fine-tune:
     • music volume
     • SFX
     • voiceover
   • Add/change characters
   • Adjust pronunciation of words

Scene-level generative replacement

• For a specific shot, request “generate new media”
• Describe the desired clip and choose clip length

Export/download

• Once satisfied, click the interface button to download the video.

---

Key Tutorial Outputs / Claims

• Generate 10-minute AI videos without stitching many short clips.
• Use prebuilt workflows so users don’t need to write prompts.
• Use high-end/selected models during media generation.
• Edit efficiently via:
  • scene replacement
  • script updates
  • language translation
  • settings controls

---

Main Speakers / Sources (as Indicated by Subtitles)

• Primary speaker: the single narrator/creator who demonstrates Invideo and repeatedly references their own actions (“I”).
• Product/tool source: Invideo (also referenced as “Nvido”) plus the workflow/model names mentioned by the speaker (e.g., Google “VO 3.1” and Sora 2).