Apex Studio

Effortless Open-Source Content Creation.

Apex was built around a simple belief: using free and open-source diffusion models should feel effortless. Making the model run shouldn’t be the challenge, the challenge should be the ambition, depth, and complexity of the content you choose to create.

📦 Download & Install

The fastest and easiest way to use Apex Studio is to download the latest release for your operating system. For users who prefer a pre-packaged backend (API and Python environment), you can find optimized bundles on Hugging Face.

For other versions or to see the changelog, visit our GitHub Releases.

---

🛠️ Building from Source

If you are a developer and want to contribute or run the latest development version, please refer to our Building from Source Guide.

Features

---

• Video editor built with JavaScript and Electron for easy video creation, and timeline based editing.
• Easy preprocessing by dragging your desired preprocessor model onto a media clip.
• Point-based masking with positive and negative markers for precise, intuitive control over exactly what gets masked.
• Use any LoRA from popular model hubs like Hugging Face or Civitai—bring your own checkpoints, styles, and character LoRAs and plug them directly into your workflow.
• Projects are saved as simple JSON files, making them easy to version, hand-edit and share
• No node graphs: projects stay straightforward and portable, improving cross-compatibility across machines, setups, and collaborators.
• Built-in queueing so you can line up multiple renders/generations and let Apex process them in order.
• Denoised latent previews at predetermined intervals, so you can watch generations evolve as they render.
• Built-in video controls including speed changes, frame interpolation, and keyframe selection.
• Editing controls for trimming, slicing, cropping, and much more.
• Hundreds of effects available to use within your creation.
• Audio controls including detaching audio from video, waveform manipulation, noise reduction, and more.

Models

The canonical list of shipped manifests lives under apps/api/manifest/ (this README links to the current set).

• Image Models

  • Chroma HD Text To Image
  • Flux Dev Fill
  • Flux Dev Kontext
  • Flux Dev Text To Image
  • Flux Krea Text To Image
  • Flux2 Dev Text To Image Edit
  • Flux2 Dev Text To Image Edit Turbo
  • Flux2 Klein 4B Text To Image Edit
  • Flux2 Klein 9B Text To Image Edit
  • Nunchaku Flux Dev Kontext
  • Nunchaku Flux Dev Text To Image
  • Nunchaku Flux Krea Text To Image
  • Nunchaku QwenImage Edit 2509 Lightning 8 Steps
  • Nunchaku QwenImage Edit Lightning 8 Steps
  • Nunchaku QwenImage Lightning 8 Steps
  • QwenImage
  • QwenImage 2512
  • QwenImage Edit
  • QwenImage Edit 2509
  • QwenImage Edit 2511
  • SRPO Text To Image
  • Wan 2.2 A14B Text To Image
  • ZImage
  • ZImage Turbo
  • ZImage Turbo Control

• Video Models

  • HUMO 17B
  • HunyuanVideo 1.5 I2V
  • HunyuanVideo 1.5 T2V
  • OVI 10B 10S
  • OVI 10B 5S
  • Wan 2.1 14B Image To Video 480P
  • Wan 2.1 14B InfiniTalk
  • Wan 2.1 14B MultiTalk
  • Wan 2.1 14B SCAIL
  • Wan 2.1 14B Text To Video
  • Wan 2.1 14B VACE Control
  • Wan 2.1 14B VACE Expand Swap
  • Wan 2.1 14B VACE Painting
  • Wan 2.1 14B VACE Reference To Video
  • Wan 2.2 14B Animate
  • Wan 2.2 14B Speech To Video
  • Wan 2.2 5B Text To Image To Video Turbo
  • Wan 2.2 A14B Text To Video
  • Wan 2.2 Fun A14B Control
  • Wan 2.2 Fun A14B VACE
  • Wan Lynx 14B
  • Wan Lynx Lite 14B
  • Wan 2.2 A14B First Frame Last Frame
  • Wan 2.2 A14B Image To Video
  • Wan 2.2 A14B Stable Video Infinity
  • LTX 2.0 TI2V Distilled
  • LTX 2.0 TI2V
  • MOVA Image to Video

• Upscalers

  • FlashVSR 1.1 Full
  • FlashVSR 1.1 Tiny
  • SeedVR2 3B
  • SeedVR2 7B

• Preprocessors

  • See apps/api/manifest/preprocessor/ for the built-in preprocessing model manifests (canny/depth/pose/segmentation/etc).