AI video, images, 3D and sound. Without leaving Resolve.

No. Both editions are in the download and use the same license key. Resolve Studio gets the plugin: a panel inside Resolve under Workspace ▸ Workflow Integrations that reads the clip under your playhead directly. Free Resolve gets the standalone app: its own window beside Resolve with the same apps and models, linked to your timeline through a built-in bridge. Press Connect to DaVinci Resolve once and every result can go straight to your timeline or Media Pool.

On the free Resolve, use version 21.0 (21.0.4 is the latest). Blackmagic turned off external scripting in the free 21.1, so no app can link to your timeline there: you would have to drag every result in by hand. If you already updated, install the free 21.0.4 from blackmagicdesign.com; your projects stay as they are. Resolve Studio works on every version, 21.1 included.

Own Studio? Install both. They share settings, and running both counts as one computer on your license.

New in 3.2.0

• Storyboard, a new app: a script becomes shot cards, the cast keeps its look from frame to frame, and the board plays as an animatic
• Motion graphics, its own app: 18 presets, or a chat with Claude or ChatGPT that writes the graphic or builds it in Fusion
• Graphics for your edit, on Resolve Studio: Resolve listens, your AI makes the set, and each graphic lands at its moment. Captions are Text+ clips you can edit
• An assistant that does the whole job: on Resolve Studio, a brief and a footage folder become an edit, and you can keep typing while it works
• It reads client notes from Excel, Word, PDF and Google Docs links, and never takes over your mouse or keyboard
• Your own LoRAs on every ComfyUI image and video model, and Wan 2.2 14B on your own GPU
• Object, on ComfyUI and Comfy Cloud: click, track, remove, replace or edit, then place it over your clip, frame for frame
• One panel for camera, light and move, with your own 3D light setup
• Gemma 4 as a free brain on your own computer
• New models: Kling 4.0, FLUX 3 Max, FLUX 3, Nano Banana 2.1, Luma Ray 3.2, Genjutsu Restyle and Ads Studio
• Hide an engine you do not use, and see what each result cost in Generation info

New in 3.1.9

• Higgsfield, a third cloud engine
• Talking character in Sound Design
• Elements: your characters, objects and places, called with @
• Claude and Codex connect in one press
• One Approvals switch
• Drag, drop and paste from Resolve into any picture spot

Updates are free for life. The app tells you when one is ready: Settings ▸ Check for updates.

All three. Windows and Mac get the plugin and the app. Linux gets the standalone app, because Blackmagic does not allow panels there. Apple Silicon and Intel Macs run the same universal build.

One exception: DaVinci Resolve from the Mac App Store cannot run scripts at all, so nothing can connect to it. Install the free Resolve from blackmagicdesign.com instead. Your projects are untouched.

Cloud renders run on your own fal.ai or Higgsfield key. You pay them per render, at their own price shown before you press Generate, and we take no cut. A Higgsfield render that fails or is refused is not charged. The estimate sits on the card before you press Generate, quoted from your actual clip, with the 4K finish as its own line. You see the exact charge in your fal dashboard every time.

• Images — usually a few cents each.
• Video — priced per second of output, so a short clip is typically tens of cents and a long 1080p clip on a top model costs more.
• Your own GPU — free and unlimited. Comfy Cloud is billed in credits per render.

Pick This computer in Settings ▸ Engine. The plugin finds ComfyUI on any port, starts it if it is not running, and shows only the models your machine can run. Downloads live in Settings ▸ Local models, one click each, with each model’s size. You never build a node graph.

• 8 GB VRAM — SDXL, Z-Image, Wan 2.2 5B video at 720p
• 12 GB+ — FLUX.1, LTX 2.3 at 720p
• 16 GB+ — FLUX.2 Klein, Qwen-Image at full speed, LTX at 1080p
• 24 GB (3090 / 4090 / 5090) — everything, including LTX native 1080p with audio and Wan 14B video-to-video

32 GB of system RAM is comfortable, 64 GB for the largest models. Apple Silicon is supported: the Mac build of each model downloads automatically. No big GPU? Point it at another machine (a PC at home, RunPod or Vast.ai), or use Comfy Cloud, which runs the same open models on hosted hardware.

Each local model has its own Settings: seed, steps, guidance, sampler and more. See every model’s size.

Yes, on ComfyUI, on this computer or another machine. Open a model’s Settings, press + Add a LoRA, and pick one ComfyUI already has, or bring a .safetensors file in from your computer, or drop it on the panel. With ComfyUI on another machine, put the file in that machine’s loras folder first. Set its strength, and Add to prompt puts its trigger words in. Each image and video model holds up to four, and a high and low noise pair for Wan goes on the right experts by itself.

Cloud via fal.ai: Seedance 2.5 and 2.0, Kling 4.0, Kling 3.0 Pro and 4K, Kling O1, FLUX 3 Max, MiniMax H3 and H3 Max, LTX 2.3 and 2.5, Veo 3.1, Gemini Omni Flash, Luma Ray 3.2, GPT Image 2, Nano Banana 2.1, 2 and Pro, FLUX 3, Qwen Image 3, Seedream 5, Tripo H3.1, Meshy 7.1, Rodin Gen-2.5, ElevenLabs, Stable Audio 3, MiniMax Music, Lyria 3, talking-character models and more. Cloud via Higgsfield: Cinema Studio 4.0, Soul, Marketing Studio, Grok Imagine and Higgsfield’s own versions of Kling, Seedance, Wan and MiniMax. Free on your GPU: MiniMax H3 and H3 Omni Ref, Wan 2.2 and 2.2 14B, LTX 2.3 and 2.5, FLUX.2 and FLUX.1, Qwen-Image 2512, Z-Image, SDXL and Hunyuan3D 2.0, plus ACE-Step 1.5 XL, MiniMax Music 3 and Stable Audio 3 for sound effects and music. Comfy Cloud adds Veo 3.1, Kling v3 Omni, Ideogram 4 and more. See the roster. New models are added as they release, and Lifetime includes every update.

A drawer inside every screen, with two tabs. Chat: ask it to set up a shot, or pick the cheapest model for a 4K vertical. On Resolve Studio it also builds directly in Resolve: cuts, captions, Fusion graphics, grades and render jobs. Give it a brief and a footage folder and it builds the whole edit, with a plan you can watch, and you can keep typing while it works. Presets: 32 ready-made jobs, from cutting the silences to a colour look. Make a few choices and press Run.

Settings ▸ Approvals decides whether it asks before every change or only before spending money. Generating an image or a video always waits for your press, and Undo puts your project back. It never takes over your mouse or keyboard: it works through Resolve’s own tools, and tells you the clicks for what those cannot reach.

Brains: your ChatGPT or Claude subscription (no API key), the free NVIDIA brain, your own API key, or Gemma 4 on your own computer through Ollama. Claude Desktop, Claude Code and Codex connect in one press from Settings ▸ Other apps. How the assistant works.

Solo activates on 2 computers, Studio on 5, Facility on 15. Upgrading later keeps your key: the seat count simply grows. The plugin and the app on the same machine count once.

Instantly. Your license key, the downloads and an install video land in your email the moment payment clears. Updates arrive inside the app: Settings ▸ Check for updates. Not happy within 14 days? Money back.