How to Edit VidNo's AI-Generated Scripts Before Rendering

VidNo generates scripts automatically, but you are not locked into the AI's first draft. Every script can be reviewed, edited, and re-rendered before you publish. This is the recommended workflow for anyone who wants full control over what their video says.

The Script File

After processing a recording, VidNo saves the generated script alongside the output videos:

~/recordings/
  session.mp4              # Your original recording
  session.vidno            # VidNo project file
  session_script.json      # The generated script
  session_full.mp4         # Full tutorial output
  session_recap.mp4        # Quick recap output
  session_highlight.mp4    # Highlight reel output

The script file is a JSON document with a clear structure:

{
  "metadata": {
    "title": "Building a Pagination API with Express",
    "description": "Step-by-step guide to...",
    "total_duration": "12:30",
    "chapters": 5
  },
  "chapters": [
    {
      "title": "Project Setup",
      "timestamp": "00:00",
      "narration": "We're starting with a fresh Express project...",
      "scenes": [1, 2],
      "duration": "02:15"
    },
    {
      "title": "Database Schema",
      "timestamp": "02:15",
      "narration": "For pagination to work efficiently...",
      "scenes": [3, 4, 5],
      "duration": "03:40"
    }
  ]
}

Editing the Script

Open the script file in any text editor. You can modify:

• Narration text: Fix inaccuracies, add context, remove redundancy, adjust tone
• Chapter titles: Make them more descriptive or keyword-rich for YouTube
• Chapter ordering: Rearrange sections for better pedagogical flow
• Metadata: Title and description for YouTube upload

VidNo also provides a CLI editor for quick changes:

# Open the interactive script editor
vidno script edit session.vidno

# This opens a TUI (terminal UI) with:
# - Chapter list on the left
# - Narration text on the right
# - Scene preview thumbnails at the bottom

Common Edits

Fixing Technical Inaccuracies

The AI sometimes misidentifies intent. If you extracted a utility function but Claude describes it as "creating a new service layer," edit the narration to correct it. The code context is right -- the interpretation might be slightly off.

Adding Context

Claude does not know your project history. If this is part 3 of a series, add a reference: "In the previous video, we set up the database. Now we're building the API layer on top of it."

Removing Redundancy

Claude sometimes over-explains simple operations. If it narrates "now let's import the express module from the express package" -- trim that to "importing express" or remove it entirely.

Adjusting Tone

If the script is too formal or too casual for your audience, edit the language. Some developers prefer a dry, matter-of-fact tone. Others want more personality. The AI gives you a solid draft; you shape the voice.

Re-Rendering After Edits

Once you are satisfied with the script, re-render:

vidno render session.vidno

This skips the analysis and script generation stages (stages 1-4) and goes directly to voice synthesis and video rendering (stages 5-6). A re-render typically takes 1-2 minutes.

You can iterate as many times as you want. Edit, re-render, review. The turnaround is fast enough that you can fine-tune in real time.

Workflow Recommendations

Not every video needs script editing. Here is a practical framework:

• Quick content (shorts, recaps): Skip editing. AI output is good enough for short-form content.
• Standard tutorials: Do a quick read-through. Fix obvious issues but do not over-edit. 5 minutes of review.
• High-stakes content (conference talks, course material): Full script review and editing. 15-30 minutes of polishing.

The goal is to spend your time on creative decisions (what to say, how to frame concepts) rather than production work (cutting footage, syncing audio). VidNo handles production. You handle editorial.