ComfyDirector is an experimental open-source playground on top of ComfyUI. Give it a prompt, pick a workflow template, keep continuity references close, and try ideas without learning node graphs, wiring models by hand, or treating every output like a finished production.
Switch between cinematic, widescreen, portrait, and social crops to see how framing changes the result. It is a quick way to test composition before you commit to a workflow.
ComfyDirector started as a way to mess with AI video without constantly wrestling with raw node graphs.
It wraps ComfyUI in a simpler experimental layer for storyboard review, template routing, continuity references, refinement, audio, and assembly. The point is not to pretend this is polished studio software. The point is to make it easier to play, compare, rerun, and learn what works.
A lighter way to mess around with prompts, templates, references, refinement, audio, and workflows without learning raw ComfyUI first.
The Library lets you save characters, props, locations, styles, and motion notes, link them into projects, and bind them to scenes. Supported templates can route a primary character image directly into image generation.
Start with a rough idea and let the built-in LLM turn it into scenes, shot directions, and prompts you can actually test.
See scene beats, images, reference usage, and dialogue choices before you commit to heavier video runs.
Promote useful images into reusable references, curate them globally, and bind characters or other continuity notes back into scenes.
Use opt-in image refinement, manual masks, and assist flows to repair or redirect shots before sending them into video generation.
Swap between cinematic, widescreen, portrait, and social framing to see how the same idea changes.
Prompts, images, videos, audio, workflow exports, reference metadata, and rerun history stay attached to each project.
Make a rough cut inside the app. Reorder clips, trim them, add transitions, adjust SFX mix, and queue optional movie mastering.
Write the idea in plain English. No node graph needed.
AI turns it into scenes you can review, tweak, and approve before the heavy stuff runs.
Pick a template and let ComfyDirector run the matching ComfyUI workflow behind the scenes.
Clips and audio can be stitched into a simple final video.
Built for messing around: rerun stages, swap templates, tweak prompts, change seeds, and keep going until something interesting shows up.
11 Gen 2 templates across 5 GPU tiers — from lighter 8GB experiments to external/self-hosted LTX 2.3 cinematic, I2V, and lip-sync workflows.
Fast text-to-video with Wan 2.2 5B. Sound design via MMAudio. Perfect for quick experiments on modest GPUs.
Synchronized audio and video in a single pass with LTX-2. Camera control, face enhancement, and LoRA support built in.
Dual-expert MoE architecture delivers the highest quality text-to-video. LoRA support for style customization.
Specialized anime pipeline with curated LoRAs: retro 90s, modern HD, and Kawajiri styles. Built on the 14B MoE backbone.
The complete cinematic pipeline. Hybrid video (S2V lip-sync + LTX-2 motion), TTS narration, multi-character support, face enhancement, and opt-in image refinement before video.
Experimental LTX 2.3 native audio+video. Keeps the Qwen image and camera stages, then upgrades the final pass to newer synchronized dialogue and ambient audio generation.
Direct LTX 2.3 text-to-video with synchronized audio. Best when you want the newest cinematic backend without the full multi-stage image and camera pipeline.
Single-image LTX 2.3 image-to-video with storyboard image upload fallback, native synchronized audio, and two-pass cinematic refinement.
Guided LTX 2.3 image-to-video with up to five visual anchors per clip through the WhatDreamsCost sequencer workflow pattern.
AITold-style cinematic lip-sync from one scene image and uploaded or generated dialogue audio, with guide-audio separation and two-pass refinement.
Routed LTX 2.3 ID-LoRA talking-head clips from storyboard images plus uploaded or generated dialogue, with Qwen or ZImage fallback.
Plus 6 legacy templates (Gen 1) still available for backward compatibility.
ComfyDirector runs locally if you have enough GPU headroom. The lighter templates start
around 8GB VRAM. The newest LTX 2.3 I2V and lip-sync routes are best treated as
external/self-hosted 32GB+ workflows today, while the managed Docker/WSL runtime remains
below recommendation for the heaviest LTX 2.3 paths. Most hands-on validation has been on
an RTX 5090.
If your machine is smaller, start with the lower-VRAM templates or use cloud GPUs for
heavier runs. The managed ComfyUI service runs through Docker on Linux or WSL2; if you
already have a native Windows or Linux ComfyUI install, point ComfyDirector at it as an
external runtime.
ComfyDirector is an open-source side project built in spare time. If it saves you from wrestling with
raw ComfyUI just to test a prompt, Buy Me a Coffee helps cover the time and experiments behind it.
If you would rather not donate, starring the repo, opening issues, and contributing workflows or
templates are still extremely useful signals.