Video summary

AI Videos | 02 घंटें में सब कुछ | बनाने से कमाने तक का सफ़र

Main summary

Key takeaways

Educational

Main ideas / lessons conveyed

  • AI video creation can be done quickly (in minutes) rather than months—if you follow a practical workflow: script → scenes → generate visuals → add audio → combine/export.
  • Speed + scale matters: creators who adopt new AI capabilities faster can produce and monetize more content.
  • Humans remain important: AI helps generate, but editing, correction, and final judgment are still required.
  • Quality comes from structure: short scene durations, consistent timing, and checking for mistakes—especially audio timing and sensitive visual elements like maps/flags.
  • Practical training beats theory: the speaker emphasizes hands-on demonstration, troubleshooting, and patience.
  • Monetization depends on uploading + creativity: AI-generated outputs may not monetize automatically; creators must produce upload-ready, engaging videos.

Methodology / step-by-step workflow taught (detailed)

A) Overall pipeline to create a 2D documentary / narrated video

  1. Prepare the script

    • Use ChatGPT to generate a script based on a prompt.
    • Pick a topic (example: Iran–America–Israel war geopolitical documentary).
    • Specify a target length (the speaker mentions ~5 minutes as feasible, though scenes are chunked shorter).

  2. Break the script into scenes

    • Key instruction: AI platforms limit scene duration to roughly 8–10 seconds per segment.
    • Speaker’s rule:
      • Scene length: ~10 seconds
      • Total scenes: e.g., 5 scenes for short output (later mentioned 10 scenes for longer output)

  3. Generate scene visuals

    • Move from ChatGPT to Google’s tool (referred to as “Google Flow” / “Omni” / “OmniFlash”).
    • Configure vertical vs horizontal output.
    • Select a voice / VO option (speaker mentions “VO 3.1 / OmniFlash”).
    • Use the correct prompt text for each scene and submit.

  4. Create audio narration

    • Ensure the voice matches the script.
    • If audio is missing or incorrect, the speaker advises adding/adjusting audio using commands and re-running/repairing.

  5. Combine scenes into one video

    • In the editor area (Google Flow), use a timeline-like approach:
      • Click the scene slot
      • Use plus/add clip
      • Add scene 1, scene 2, scene 3, …
    • Verify timing because issues like repeated beats / repeated audio lines can happen.

  6. Export / download

    • Download/export after verifying the combined video (speaker demonstrates an assembled output of about ~50 seconds).

  7. Quality checks

    • Audio presence: sometimes AI “leaves out audio”; fix by re-adding audio properly.
    • Timing mistakes: example—“heartbeat beats twice” and needs correction.
    • Visual mistakes: brightness/color issues and missing/incorrect visuals.
    • Sensitive content: check maps/flags for correctness to avoid objectionable representations.
    • Credits consumption: avoid changing too many settings/outputs at once to prevent wasting credits.

B) Using an “Agent” feature to generate multiple scenes faster

  • The speaker demonstrates an agent workflow that:
    • Generates a batch of scenes (example: scene 6 to scene 10)
    • Reduces constant manual switching
    • Lowers the chance of mistakes caused by manual steps
  • Costs: agent creation can consume a large credit chunk (speaker mentions ~150 credits for five scenes).
  • Behavior: it may “stick” temporarily; speaker advises patience while settings settle.
  • Output: returns a sequence of generated scenes ready for combine/export.

C) Creating an avatar (Gemini / Google avatar workflow)

  1. Open Google Gemini
  2. Go to Avatar settings
  3. Scan using a phone camera
    • Enable airplane mode
    • Scan via Google’s camera/Lens-like flow
    • A link appears after scanning; confirm to proceed
  4. Capture voice via guided prompts
    • The system shows numbers; user must speak numbers quickly
    • System may ask for head turns/pose changes; user follows prompts
  5. Generate avatar video
    • Use an “original window”/avatar section to generate outputs
    • Speaker says it works best in the primary Gemini window

Constraints / content safety

  • Avoid copyright music
  • Voice matching may vary slightly depending on avatar refinement

Practical monetization guidance (as stated)

  • Upload-ready AI videos can be monetized only if the creator:
    • uploads the videos, and
    • adds enough creativity/quality to meet platform expectations.
  • The speaker discourages “wasting time” searching elsewhere; he claims to provide direct instructions and templates.

Free credits method and renewal (as explained)

How to get “free first month” credits (Gemini subscription)

  1. Open Google Gemini
  2. Go to Settings → Subscriptions
  3. Choose the Monthly option
  4. First-time users may see ₹0 for the first month (if logged in with a new email).
  5. Payment mechanism described:
    • Deduct a small amount (speaker mentions ₹2) via payment/auto-pay setup to activate access
    • After the month, the full subscription price would be charged unless auto-pay is canceled

Auto-pay cancellation

  • Cancel auto-pay a bit before the month ends (example: after 25 days / set an alarm).
  • If canceled in time, the monthly charge should not occur again.

How credits work in Google Flow

  • After activating Gemini, credits appear in Google Flow.
  • Speaker examples:
    • ~50 credits daily
    • Also mentions about 1000 free credits (later reiterated as practice capacity)
  • Advice:
    • Monitor remaining credits
    • If credits end, create/use another ID (speaker says not to waste credits)

Where to find resources (from the video)

  • Speaker claims to provide:
    • Links in the video description
    • WhatsApp group link
    • Telegram link
  • He also claims the prompt/workflow/templates are available in description/support channels.

Speakers / sources featured

Speakers

  • Unnamed main speaker: appears to be the instructor/host; repeatedly addresses viewers as “brother” and demonstrates screen-based steps.
  • Ruhi: AI-generated avatar voice used in demonstrations.
  • Amandeep Sir: referenced as the host’s persona inside the avatar script.

Tools / platforms / sources mentioned

  • ChatGPT
  • Google Gemini
  • Google Flow
  • Omni / OmniFlash (model/tool naming as stated)
  • Google services for avatar/scanning (phone scanning via camera/Lens-like flow)
  • VO 3.1 (voice option referenced)
  • WhatsApp (group link)
  • Telegram (link)

Original video