Skip to content

ardahash/Autovideo_v2

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

15 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

title description
Autovideov2
Streamlit app that turns a topic or script into a narrated video with optional AI visuals (Pexels or Pixar-style AI images), narrator cut-ins, background music, and a final review step.

I made this a dark documentary tone, but you can modify it using the code.

🕯️ Autovideo (v2)

A Streamlit app that turns a topic or your own script into a dark, ominous narrated video with:

  • AI script generation (optional)
  • Premium voiceover via ElevenLabs (cached)
  • Visual pipeline options
    • v1: Copyright-free stock visuals from Pexels (video + images)
    • v2: Pixar-style AI images generated per script segment (cost-controlled)
  • Final manual visual review step (replace bad clips before rendering)
  • Optional background music
  • Optional narrator face cut-ins (every ~minute)

Features

✅ Script

  • Upload .txt / .md scripts
  • Or generate scripts from a topic (dark documentary tone)

✅ Voiceover

  • ElevenLabs TTS (cached using a hash of the script)
  • Automatically regenerates if cached file is corrupted / too small

✅ Visuals (choose one)

Option A — Pexels slideshow (v1)

  • Splits script into segments
  • Uses a small OpenAI model to generate a short search query per segment
  • Fetches Pexels video first, falls back to Pexels image
  • Lets you:
    • edit the query per segment
    • replace clips with one click
    • preview everything before render

Option B — Pixar-style AI frames (v2)

  • Generates stylized scenes aligned with the script
  • Creates a cinematic slideshow with audio
  • Designed for brand-unique visuals without stock footage

⚠️ Video generation (true animated clips) is not enabled in this repo yet — v2 currently uses AI images stitched into a slideshow.


Project Structure

  • appv2.py — main Streamlit app
  • assets/
    • background.jpg (optional default)
    • music.mp3 (optional default)
  • output/
    • cached voiceovers: voice_<hash>.mp3
    • downloaded Pexels media: frame_###.mp4 / frame_###.jpg
    • generated AI images: pixar_scene_###_<hash>.png (v2)
    • final videos: <topic>_<hash>.mp4

Requirements

  • Python 3.10+ recommended (3.11 works great)
  • FFmpeg installed (required by MoviePy)
  • API keys:
    • OPENAI_API_KEY (optional but recommended)
    • ELEVENLABS_API_KEY (required)
    • PEXELS_API_KEY (required only for Pexels mode)

Run

From PowerShell or terminal in your project folder:

py -m venv .venv
.\.venv\Scripts\activate
py -m pip install --upgrade pip
py -m pip install -r requirements.txt
you can do --legacy-peer-deps if errors arise