AI Movie Maker

AI Movie Maker is a modular Python application anchored by the ai_movie_maker.py entrypoint and a supporting aimm/ package. It combines a command line interface and a Gradio GUI to orchestrate AI-assisted storyboarding, animation, and video production workflows. The tool coordinates Stable Diffusion / AnimateDiff rendering through ComfyUI, automates voiceovers and captions, keeps project state in SQLite, and exports final edits with moviepy-powered transitions. This repository also contains a Windows installer definition and helper scripts for dependency bootstrapping.

Key Capabilities

• Hybrid CLI + GUI – Drive the entire workflow from scripted commands or through a multi-tab Gradio interface that covers setup, scripting, keyframe planning, shot controls, audio tools, rendering, review, and collaboration.
• Project management – Each project lives inside ~/.ai_movie_maker/projects/<name> with scenes, shots, keyframes, assets, and renders tracked in project.db (SQLite). CLI verbs and GUI actions both keep the database in sync.
• Rendering pipelines – Support for AnimateDiff (via ComfyUI queue API), Ken Burns-style moves on stills, and a RunwayML-style cloud stub. Shot-level width, height, frame count, FPS, prompts, and transitions are editable.
• Audio automation – ElevenLabs-powered voiceovers with automatic silence trimming (pydub), speech_recognition transcription with LLM caption polishing, and a Suno music generation stub.
• Post-processing helpers – Optional RIFE frame interpolation, Real-ESRGAN upscaling hooks, transition-aware export (fade / crossfade) to MP4, GIF, or AVI, and ffmpeg utilities for additional processing.
• Collaboration stubs – Quick share-link and project ZIP helpers to coordinate work through file-sharing services.

Repository Layout

Path	Description
ai_movie_maker.py	Lightweight entry script that dispatches to the CLI.
aimm/	Package containing modular components (core.py, gui.py, cli.py, and helpers).
requirements.txt	Python packages required by the app and installer.
setup_models.py	Hugging Face download helper invoked during workspace setup.
installer.iss	Inno Setup script that produces AI_Movie_Maker_Setup.exe on Windows.

Installation

Windows (recommended)

1. Install Inno Setup 6.2+.
2. Clone this repository and review the binary download URLs and SHA256 hashes defined near the top of installer.iss.
3. Compile the installer via Build → Compile inside the Inno Setup Compiler. The resulting AI_Movie_Maker_Setup.exe appears in Output.
4. Run the installer. It downloads FFmpeg, RIFE, Real-ESRGAN, wkhtmltopdf, ComfyUI Portable, installs Python packages into ComfyUI’s embedded interpreter, and creates Start Menu / Desktop shortcuts alongside a run.bat launcher.
5. Re-run the installer with /SILENT for unattended deployments. Optional components (interpolation, upscaling, wkhtmltopdf, model sets) can be toggled through installer checkboxes or preseeded via silent-mode parameters.

Linux / macOS

1. Ensure python3, pip, git, curl, and unzip are available.
2. Create and activate a virtual environment, then install dependencies: pip install -r requirements.txt.
3. Place ComfyUI inside ~/.ai_movie_maker/bin/ComfyUI (or adjust the path in the script) and install its requirements.
4. Run python setup_models.py to fetch default Stable Diffusion / AnimateDiff checkpoints via huggingface_hub.
5. Launch the app with python ai_movie_maker.py init followed by python ai_movie_maker.py gui.

Quick Start

1. Initialise workspace: python ai_movie_maker.py init
2. Create a project: python ai_movie_maker.py create my_movie
3. Generate keyframes: python ai_movie_maker.py gen-keyframes my_movie
4. Render a shot: python ai_movie_maker.py run my_movie shot:1
5. Export final video: python ai_movie_maker.py export my_movie mp4
6. Launch GUI: python ai_movie_maker.py gui

All commands use secure key storage via keyring. Set API keys inside the GUI Settings tab or through your OS keychain utility before invoking LLM or ElevenLabs features.

Gradio GUI Overview

The GUI organises tasks into dedicated tabs:

• Checklist – Run dependency checks (FFmpeg, RIFE, Real-ESRGAN, ComfyUI).
• Settings – Configure API keys for OpenAI, Anthropic, and ElevenLabs.
• Script / Storyboard – Edit the script, sync scenes/shots, and manage prompts.
• Shot Editor – Update shot metadata, camera settings, transitions, and assets.
• Audio / Captions / Music – Trigger voiceover generation, auto-captioning, and music stubs.
• Render – Submit renders (AnimateDiff, Ken Burns, cloud stub) and preview outputs.
• Review & Export – Combine shots with transitions and export to MP4/GIF/AVI.
• Collaboration – Save or load project ZIPs and retrieve share links.
• Help – Quick usage reminders and support resources.

CLI Reference

Command	Purpose
init	Create workspace directories and sentinel file.
check	Print dependency status for FFmpeg / RIFE / Real-ESRGAN / ComfyUI.
list	Display all projects found in the workspace.
create <name>	Scaffold a new project with default script content.
load <name>	Load an existing project (required before most verbs).
sync-script <name> <path>	Replace the stored script with the contents of <path>.
gen-keyframes <name>	LLM-driven keyframe prompts for every shot.
`run project	scene:
`export <mp4	gif
export-pdf <name> <path>	Produce a PDF storyboard via wkhtmltopdf (if installed).
save-project <name> <zip>	Zip the project folder for sharing/backups.
load-project <zip>	Restore a project from a ZIP archive.
storyboard <name>	Print a text storyboard summarising scenes and shots.
edit-shot <name> <shot_id> <field> <value>	Update shot metadata (width, height, fps, etc.).
voiceover <name> <shot_id> <text...>	Generate an ElevenLabs voiceover for a shot.
captions <name> <shot_id>	Transcribe + refine captions for the associated voiceover.
music-stub <name> <scene>	Produce a Suno placeholder music clip for a scene.
share <name>	Return a collaboration stub link.
fetch-share <url>	Demonstrate retrieving a shared project link.
gui	Launch the Gradio interface.

Additional helpers include run my_movie shot:3 (shot aliases), run my_movie scene:1 (batch render a scene), audio utilities, music stubs, and share-link retrieval. Use python ai_movie_maker.py --help to view the latest command list.

Frequently Asked Questions

Do I need a GPU?
No. AnimateDiff runs faster on NVIDIA GPUs, but the script falls back to CPU-friendly modes (Ken Burns, cloud stub) if GPU-only dependencies are unavailable.

What happens if LLM APIs are unreachable?
Keyframe planning and caption refinement gracefully fall back to the shot description so you can continue working offline.

Where are projects stored?
Projects live under ~/.ai_movie_maker/projects/. Each project folder contains assets/, renders/, and a project.db file.

Can I use custom models?
Yes. Place additional checkpoints under ~/.ai_movie_maker/models/ and update ComfyUI workflows or CLI parameters as needed. The installer exposes a custom model path option for Windows builds.

How do I update binaries?
Re-run the Windows installer or re-fetch the binaries manually and replace the contents of ~/.ai_movie_maker/bin/. The check command reports missing components.

Is there a way to collaborate?
Use the Collaboration tab or the save-project / load-project commands to exchange ZIP archives via Dropbox, Google Drive, or similar services. A stub share-link helper is included for quick handoffs.

Where are logs stored?
CLI output prints to stdout/stderr. ComfyUI logs live under ~/.ai_movie_maker/bin/ComfyUI/logs/, and FFmpeg errors are surfaced through gradio alerts or CLI messages.

Support & Contributions

Please open issues or pull requests if you encounter bugs, need installer tweaks, or want to contribute enhancements. Include detailed logs and reproduction steps to help triage quickly.