How to use the AI Video Studio DaVinci Resolve plugin for cloud and local AI

01 Sep 2026 04:37 11,797 views
A new DaVinci Resolve plugin, AI Video Studio, brings text-to-video, image generation, relighting, rotoscoping, 3D tools, and an AI assistant directly into your timeline. It works with both pay-per-use cloud models and free local models on your own GPU, so you can ditch monthly AI subscriptions and keep everything inside Resolve.

AI is rapidly changing how editors, VFX artists, and 3D creators work, but the tools only become truly powerful when they fit neatly into your existing workflow. That’s exactly what the AI Video Studio plugin for DaVinci Resolve is designed to do: bring cutting-edge AI models directly into Resolve, without browser tabs, manual uploads, or expensive monthly subscriptions.

What is AI Video Studio for DaVinci Resolve?

AI Video Studio is a workflow integration for DaVinci Resolve Studio that embeds a full AI app inside the Resolve interface. It connects directly to your timeline, lets you generate and edit media with the latest AI models, and drops the results straight into your project—no import/export steps required.

The plugin supports both cloud-based models and free local models running on your own GPU. You can pay per render via cloud APIs, use hosted ComfyUI GPUs, or run everything locally for zero ongoing cost (beyond your hardware and electricity).

How to open the plugin inside DaVinci Resolve

Once installed on DaVinci Resolve Studio, you open the plugin from the top menu:

Workspace → Workflow Integrations → AI Video Studio

This brings up the main interface. You can dock the window to the side of the Resolve UI and pin it so it stays visible while you edit, making it feel like a native part of your workflow.

AI video generation: text to video and video to video

The plugin includes a dedicated video generation section where you can create clips from scratch or transform existing footage with state-of-the-art models.

Supported models and modes

AI Video Studio supports a range of video models, including:

C-DiT / C-Dance 2.0 & 2.5 (high-end, more expensive models)
Minimax and other more affordable or local-friendly models

You can use:

Text-to-video – generate a new clip from a written prompt
Image-to-video – animate a still image
Video-to-video – restyle or relight an existing shot while preserving motion

Camera, lens, and lighting presets

Instead of over-prompting every detail, the plugin gives you structured controls:

Camera & lens presets – choose focal length, lens type, and camera look
Lighting presets – studio, interior, exterior setups, plus a custom lighting prompt field
Light rig controls – fine-tune the direction and style of light in your scene
Camera animation presets – built-in moves so you don’t have to describe motion in text

Transparent pricing before you render

For cloud models, the plugin shows an estimated cost before you generate:

• Example: C-Dance 2.5 ~ $2 for a 5-second 720p clip
• Example: Minimax ~ around $1 for 5 seconds at 4K

Pricing is displayed per model and at the bottom of the interface, so you always know what you’re spending before you render.

Editing your existing footage with AI

The Edit Video tab connects directly to your Resolve timeline. When you select a clip in the timeline, the plugin reads that frame and lets you apply AI transformations.

For example, you can:

• Choose a model like C-Dance 2.0
• Type an instruction such as: “Make this a nighttime scene and remove the microphone and arm”
• Optionally add reference images, video, audio, or even a screenshot of the current frame

After generating, the result appears in the plugin’s Recent panel. You can then drag the new clip directly into your timeline or media pool. The original shot stays intact, and the AI version becomes just another asset in your project.

Specialized tools: drawing, reframing, relighting, and more

Beyond basic generation, AI Video Studio includes a set of focused tools tailored to real post-production tasks.

Draw to video

This mode lets you draw over an image or frame to guide changes—useful for targeted edits, roughing in shapes, or indicating where you want alterations to appear in the final video.

Smart reframing

The Reframe tool automatically converts your footage into other aspect ratios, like 9:16 for vertical content. AI helps keep important subjects centered and framed correctly, which is ideal for repurposing horizontal edits for social platforms.

Relighting entire scenes

With the Relight preset, you can generate reference frames and apply new lighting looks across a whole video. This can simulate different times of day or moods without complex 3D relighting or heavy grading tricks.

AI upscaling

Built-in upscalers let you increase resolution while preserving or enhancing detail. This is handy for turning test renders or lower-resolution outputs into higher-quality deliverables.

Automatic rotoscoping

The plugin also includes an automatic rotoscope function to create quick mattes of your subject:

• Choose a mode like “dynamic” and a resolution (e.g., 2K)
• Select “cut out and matte”
• Generate to get both a cutout and a separate matte clip

The masks are surprisingly clean considering the speed, which makes this great for fast composites, stylized effects, or quick isolation work.

Image generation and VFX backdrops

AI Video Studio isn’t just for video. It also includes image-focused tools that are especially useful for VFX workflows.

Image generation

You can generate still images directly inside Resolve using modern image models. These can be used for concept art, thumbnails, matte paintings, or as reference material for video transformations.

Backdrop generator for green screen

The Backdrop Generator is designed to create background plates for green-screen shots. You can:

• Choose from themed options and presets
• Provide reference images
• Generate a matching background to key behind your subject

This is especially useful if you’re building virtual sets or need quick, consistent environments without a full 3D pipeline.

Generating 3D models with AI

For 3D artists, the plugin offers a 3D Model Generator that uses tools like Tripo and Meshy. You can feed it reference images, and it will generate a 3D model based on those inputs.

The generated models can then be brought into your 3D software of choice for further refinement, or used as quick assets for previs and concept work.

Turning raw 3D renders into photoreal shots

One of the most powerful features is the 3D & AI workflow, which is designed to take simple 3D renders and push them toward photorealism using video models like C-Dance 2.5.

The general workflow looks like this:

1. Render a basic clay or flat pass from your 3D tool (e.g., Unreal Engine 5).
2. Collect reference images that match the style, lighting, and mood you want.
3. Feed the clay render and references into AI Video Studio.
4. Let the model generate a final, stylized or photoreal shot while preserving the original motion and camera movement.

This approach can dramatically speed up rendering and look development, and it points toward a future where AI becomes a standard part of the 3D rendering pipeline. For a broader look at where DaVinci is heading with AI, you may also want to check what’s new in DaVinci Resolve 21’s AI upgrades.

AI-generated sound design

AI Video Studio also includes a sound design section built around 11 Labs. You can:

• Generate sound effects
• Create sounds that are context-aware and based on your clip

This helps you quickly build audio layers that match new AI-generated visuals without leaving Resolve or hunting through external sound libraries.

Node Studio and the built-in AI assistant

For more advanced users, the plugin includes a Node Studio area where you can chain together different tools and models in a more modular way.

There’s also a compact AI assistant panel built into the interface. You can:

• Ask for prompt ideas
• Describe what you want to create
• Use different LLMs as the backend, including local and cloud options

One notable option is NVIDIA serverless APIs, which can run for free on your own system. You can also plug in your own API keys for providers like Claude, OpenAI, DeepSeek, or even local models stored on your own machine. For example, you might ask it to “create an image of a dog,” select an image model, and generate directly inside the plugin.

Setting up cloud vs local engines

All configuration happens in the plugin’s Settings panel, accessible from the top-right of the interface. Here you choose how and where your AI jobs are processed.

1. Fall AI (pay-per-render cloud)

In Fall mode, the plugin uses the Fall AI API:

• You create an API key in the Fall AI dashboard and paste it into the plugin.
• You pay per generation—images cost cents, short clips cost around $1–$2 depending on model and resolution.
• The plugin shows the exact estimated cost before you render.

This is ideal when you want the highest-quality models and are okay with small, predictable cloud costs instead of a monthly subscription.

2. ComfyUI local (free on your GPU)

Switching the engine to ComfyUI Local lets you render for free on your own GPU:

• You need a working ComfyUI installation running in the background.
• The plugin will show when it’s connected and which GPU models are ready.
• You download and store models (like Minimax H3, LTX 2.3, Flux) on your own disk.

Requirements and notes:

• Video models typically need around 24 GB of VRAM.
• Image models can work with 8–16 GB VRAM.
• All generations are unlimited, private, and can run offline, limited only by your hardware.

You can even point the plugin to a remote ComfyUI server if you want to render on another machine (for example, a powerful desktop) while editing on a laptop.

3. Comfy Cloud (hosted GPUs with credits)

If you like the ComfyUI ecosystem but don’t want to run heavy models locally, you can use Comfy Cloud:

• You buy monthly credits (for example, starting with 400 credits).
• Credits are valid for up to one year, so they don’t disappear every month if you don’t use them.
• Jobs run on big hosted GPUs, which is useful when you need premium models like C-Dance but don’t own a high-end GPU.

Account settings and output management

Within the Accounts section of Settings, you can configure keys for:

• NVIDIA APIs (for the assistant)
• Claude, OpenAI, DeepSeek, and other LLM providers
• Local models for a fully self-hosted assistant

You can also define an output folder where all AI generations are stored. Every render lands both in your Resolve project and on disk, so you always have direct access to the source files.

Pricing model vs subscription tools like Higgsfield

Many AI video platforms use a monthly subscription model where credits reset every month, whether you use them or not. Higgsfield, for example, offers a polished web-based experience with strong tools, but plans can be around $60/month or even $130/month for heavier use, and credits typically expire monthly.

AI Video Studio takes a different approach:

• The plugin itself is a one-time purchase (via Gumroad) with no recurring subscription.
• You then choose how to pay for AI compute:

1. Fall AI – pay-per-render, no monthly commitment.
2. Comfy Cloud – top up credits that last up to a year.
3. Local GPU – free, unlimited renders on your own hardware.

If you don’t render anything in a given month, your cost is effectively $0. Nothing expires on your disk, and all outputs stay local unless you choose cloud processing.

If you are exploring other AI video workflows inside Resolve, you might also find it useful to compare this approach to the Higgsfield DaVinci Resolve plugin, which leans more on a subscription-style credit system.

Real-world usage: when to use local vs cloud

A practical way to work with AI Video Studio is to mix and match local and cloud modes:

Local models for exploration and personal projects:
– Test ideas, iterate on looks, and experiment freely.
– No per-render cost and everything stays private.

Cloud models (Fall AI or Comfy Cloud) for client work and final shots:
– Use the highest-quality models and resolutions.
– Pay only for what you actually render.
– Offload heavy jobs from your own GPU if needed.

This hybrid approach avoids the “use it or lose it” pressure of subscriptions and lets you choose the right balance of cost, quality, and speed for each project.

Platform support

AI Video Studio works with:

DaVinci Resolve Studio (paid version) on Windows and macOS
• A standalone app for Linux via a separate installer

That makes it accessible across most professional editing setups, whether you’re on a workstation, a MacBook, or a Linux-based pipeline.

Why this plugin can replace multiple AI subscriptions

By bringing video generation, video-to-video editing, image creation, green-screen backdrops, 3D model generation, photoreal 3D rendering workflows, sound design, and an AI assistant into a single Resolve-integrated tool, AI Video Studio can stand in for several separate AI subscriptions.

Because you only pay for the compute you actually use—or nothing at all when running locally—it’s an appealing option for editors and artists who want powerful AI tools without committing to another monthly bill.

Share:

Comments

No comments yet. Be the first to share your thoughts!

More in Video Editing