tin2tin/Pallaidium

What it solves

Pallaidium is a generative AI movie studio integrated directly into the Blender Video Sequence Editor (VSE). It allows filmmakers to prototype an entire production cycle—from screenplay and storyboarding to animation, voiceovers, and music—using AI before committing to actual filming. This workflow enables creators to refine their vision and extract a finalized screenplay based on an AI-generated prototype.

How it works

As a Blender add-on, Pallaidium provides a sidebar panel in the VSE where users can generate media based on text prompts or existing timeline strips. It supports a wide range of generative tasks including text-to-video, image-to-video, text-to-speech, and text-to-music.

By default, the tool runs models locally on NVIDIA GPUs, but it also supports remote backends (such as ComfyUI, fal.ai, or custom OpenAI-compatible servers) to offload computation to more powerful machines. It integrates advanced AI controls like ControlNet, IP-Adapters, and LoRAs to maintain visual consistency.

Who it’s for

It is designed for filmmakers, animators, and video editors who use Blender and want to use generative AI to rapidly prototype scenes, create placeholders, or generate final assets within their editing timeline.

Highlights

• Integrated Workflow: All AI generation happens within the Blender Video Sequence Editor.
• Multimodal Generation: Supports images, video, audio, speech, and text (transcription/screenwriting).
• Flexible Backend: Runs locally via CUDA or connects to remote servers via a provider-agnostic API contract.
• Advanced Control: Includes support for ControlNet, OpenPose, Canny, and multi-reference image inputs for high visual fidelity.
• ComfyUI Integration: Allows importing and driving ComfyUI workflows directly from Blender.