huytranvan2010/AI-auto-generate-video

Auto generate AI video with hyperframes

What it solves

This project automates the creation of 9:16 short-form videos (TikTok, Shorts, Reels) from text articles. It removes the need for manual video editing by using a deterministic pipeline that converts a script into a polished video with voiceovers and sound effects.

How it works

The system splits content creation into two parts: AI handles the script and template selection, while deterministic code handles the actual rendering.

  1. Scripting: Using Claude Code (or manually), a script.json is created containing the narration text and specific template IDs for each scene.
  2. Audio Generation: The pipeline uses OmniVoice (a local TTS engine) to generate narration for each scene and mixes in sound effects (SFX) based on keywords or scene types.
  3. Visual Rendering: HyperFrames renders HTML-based templates into MP4 clips using Chromium.
  4. Assembly: FFmpeg concatenates the clips and muxes the final audio track to produce a 1080x1920 video.

Who it’s for

Content creators who want to turn Vietnamese articles into short-form videos quickly without manual editing, and developers interested in automated video production pipelines.

Highlights

  • Deterministic Renders: The same script always produces the same video, allowing for easy iterations.
  • Coded Templates: Visuals are managed as self-contained HyperFrames projects, making them easy to customize via HTML/CSS.
  • Automated SFX: Sound effects are automatically selected based on semantic matches to the script text.
  • Local TTS: Integrates with OmniVoice for local Vietnamese text-to-speech without requiring external API keys.

Related

  • Project
    facebookresearch/mae_st

    A PyTorch implementation of Masked Autoencoders for spatiotemporal learning, enabling models to learn video representations by reconstructing masked video frames.

  • Project
    wernerturing/multi-delogo

    An application for removing logos and watermarks from videos, featuring automatic detection and the ability to handle moving logos.

  • Project
    amanchadha/iSeeBetter

    A spatio-temporal video super-resolution project that uses Recurrent-Generative Back-Projection Networks and GANs to upscale low-resolution videos while maintaining temporal consistency and fine detail.

  • Project
    Carasibana/ComfyUI-H3-FaceRefine

    ComfyUI‑H3‑FaceRefine is a custom‑node pack for ComfyUI that fixes MiniMax H3’s poor rendering of small faces in video. It detects faces, crops them to a larger canvas, runs H3 on the crops, and stitches the refined faces back. Includes auto‑ and manual‑face‑selection modes, hard‑cut detection, per‑frame denoise, and ready‑to‑use example workflows.