leamsigc/ShortsGenerator

Automate the creation of Shorts content locally with a couple simple steps.

ShortsGenerator – Automated YouTube Shorts Builder

What it does

  • Generates a short‑form video from a text prompt using AI‑written scripts, stock footage, text‑to‑speech, subtitles, background music and optional image stitching.
  • Provides a web UI (Nuxt 3) that talks to a Flask backend. The backend runs the whole pipeline and exposes a set of REST endpoints.
  • After a video is built you can schedule it to several social platforms (Instagram, TikTok, Facebook, LinkedIn, YouTube) through the MagicSync integration.

Key components

Layer Tech / Role
Backend Python 3.11, Flask, ffmpeg, Torch (for Qwen‑3 TTS), g4f/Gemini for script generation. Handles stock‑video search (Pexels API), TTS (Supertonic, TikTok, KittenTTS, Qwen‑3), subtitle rendering, video stitching, and the MagicSync API calls.
Frontend Nuxt 3 (Vue) UI, runs on localhost:3000. Lets you enter a subject, edit the generated script, pick voices, subtitles, aspect ratio, music, etc., then triggers the backend via the documented API.
AI models • Script: free g4f (browser‑cookie based) or Gemini (via GOOGLE_API_KEY).
• TTS: local Supertonic (10 voices, 33 languages), TikTok TTS (fallback), KittenTTS, Qwen‑3‑TTS (local model with 9 preset timbres, voice‑design, voice‑clone).
Media sources Stock clips from Pexels, user‑provided music files, optional image‑stitching at the start of the video.
Scheduling MagicSync – you can store multiple business accounts (API URL + token) and pick which one to use when posting.

How you use it

  1. Clone the repo, copy .env.example to .env and fill in at least PEXELS_API_KEY and IMAGEMAGICK_BINARY.
  2. Run the backend (python Backend/main.py) and the Nuxt UI (cd UI && pnpm i && npx nuxt dev --tunnel). Docker is also supported via docker compose up -d.
  3. In the web UI, type a video subject, generate a script, tweak it, choose a voice and subtitle style, add optional images or music, then click Generate.
  4. When the video finishes you can extract a thumbnail, preview it, and schedule the upload through MagicSync.

Notable features

  • Multi‑voice TTS with local high‑quality models (Qwen‑3) and fallbacks.
  • Subtitle templates (10 visual styles) generated automatically.
  • Aspect‑ratio flexibility for vertical Shorts (9:16) or other formats.
  • Hardware acceleration – detects a GPU and passes it to ffmpeg/Torch for faster encoding.
  • Multi‑business MagicSync – store separate API credentials per client and select the right one when scheduling.

Who might find it useful

  • Content creators who want to batch‑produce YouTube Shorts without manually editing video.
  • Small agencies managing several brand accounts and needing a repeatable pipeline.
  • Developers looking for an open‑source reference implementation of an end‑to‑end short‑form video generation stack.

Limitations

  • Requires Python 3.11 and a fairly heavy dependency stack (Torch, Qwen‑3 weights ≈ 3 GB).
  • Some AI services (Gemini, OpenAI) need API keys; the free g4f path depends on browser cookies.
  • Scheduling works only with the proprietary MagicSync service; the repo does not provide a generic social‑media API.

License

  • The project is released under the license found in the LICENSE file (not reproduced here).

Related

  • Project
    mushigaite/short-video-maker

    An open‑source Python tool that downloads a YouTube (or local) video, transcribes it with faster‑whisper, uses an LLM (OpenAI/Gemini/MuAPI) to rank viral‑worthy highlights, optionally adds an AI‑generated hook, and outputs ready‑to‑post vertical short clips via CLI or a tiny local web UI.

  • Project
    Norsico/Video-Materials-AutoGEN-Workstation

    An AI-powered workstation for short-video production that automates script generation, TTS voiceovers, and image asset creation to streamline content workflows.

  • Project
    naqashafzal/AI-Content-Studio

    AI Content Studio is an open‑source platform that automates the whole video‑creation workflow. A supervisory set of AI agents researches a topic, writes and QA‑checks a script, then a multimodal pipeline (Whisper transcription, OpenCV scene analysis, FFmpeg processing) produces a fully‑styled, loudness‑normalized video with captions, B‑roll and branding. It supports style‑cloning from reference clips, Urdu TTS, and 24/7 auto‑pilot campaigns, and runs via a FastAPI backend + Next.js UI.

  • Project
    YILS-LIN/short-video-factory

    An AI-powered desktop tool that automates the creation of product marketing and general content short videos from text prompts and video clips.

  • Project
    jipraks/yt-short-clipper

    YT Short Clipper v2 is a Windows‑only desktop app that automatically turns a long YouTube video into vertical (9:16) short clips. It downloads the video and its original subtitles, feeds the transcript to an LLM to pick highlight segments, cuts and re‑frames each segment, burns word‑by‑word captions, and can upload the results to Repliz. Built with React + TypeScript, Tauri (Rust), and a Python side‑car using yt‑dlp, FFmpeg, and MediaPipe, it supports any OpenAI‑compatible model or an in‑app wallet‑based AI service. The project is MIT‑licensed, beta‑stage, and Windows‑only.