Skip to main content4.1 Overview of the Editor Interface
The canvas is a visual editor where you build workflows by placing and connecting nodes.
4.2 Node Library
GMI Official Nodes
- Image, video, LLM, and audio generation nodes through API
- GMI Image Upload and Batch Images nodes
4.2.1 Video Models
Wan
- wan2.6-t2v – Text → Video: Generates videos directly from text prompts.
- wan2.6-i2v – Image + Text → Video: Animates a still image into a video.
- wan2.6-r2v – Video + Text → Video: Edits or transforms an existing video.
- wan2.5-t2v-preview – Text → Video: Fast preview version for quick testing.
- wan2.5-i2v-preview – Image + Text → Video: Preview image-to-video generation.
Veo
- veo-3.1-fast-generate-preview – Text → Video: Fast, low-latency video generation.
- veo-3.1-generate-preview – Text → Video: Higher-quality cinematic video generation.
- Veo3-Fast – Text → Video: Balanced speed and visual quality.
- Veo3 – Text → Video: Premium cinematic video generation.
Sora
- sora-2 – Text → Video: Realistic video generation from text.
- sora-2-pro – Text → Video: Higher fidelity and longer consistency videos.
Kling
- kling-v2-6 – Text → Video: Smooth, stylized text-to-video generation.
- kling-v2-5-turbo – Text → Video: Faster text-to-video generation.
- Kling-Text2Video-V1.6-Standard – Text → Video: Standard text-to-video generation.
- Kling-Text2Video-V2.1-Master – Text → Video: Advanced, high-quality video generation.
- Kling-Text2Video-V2-Master – Text → Video: Premium text-to-video model.
- Kling-Image2Video-V1.6-Standard – Image → Video: Basic image animation.
- Kling-Image2Video-V1.6-Pro – Image → Video: Higher-quality image animation.
- Kling-Image2Video-V2.1-Standard – Image → Video: Improved motion consistency.
- Kling-Image2Video-V2.1-Pro – Image → Video: Professional-grade image-to-video.
- Kling-Image2Video-V2.1-Master – Image → Video: Premium image-to-video quality.
- Kling-Image2Video-V2-Master – Image → Video: Cinematic image animation.
- kling-o1-image-to-video – Image → Video: Omni image-to-video generation.
- kling-o1-reference-to-video – Reference + Text → Video: Generates video guided by reference.
- kling-o1-edit-video – Video + Text → Video: Edits video using text instructions.
- kling-o1-flfv – Image / Video → Video: Flexible multi-input video generation.
Minimax Hailuo
- Minimax-Hailuo-02 – Text → Video: General-purpose text-to-video generation.
- Minimax-Hailuo-2.3 – Text → Video: Improved motion and visual consistency.
- Minimax-Hailuo-2.3-Fast – Text → Video: Faster, cost-efficient generation.
PixVerse
- pixverse-v5-t2v – Text → Video: Creative text-to-video generation.
- pixverse-v5-i2v – Image → Video: Image animation with visual effects.
- pixverse-v5-transition – Images / Video → Video: Generates smooth scene transitions.
Seedance
- seedance-1-0-pro-fast-251015 – Text → Video: Ultra-fast video generation.
- seedance-1-0-pro-250528 – Text → Video: Higher-quality Seedance generation.
Luma
- Luma-Ray2 – Text → Video: Cinematic video with realistic camera motion.
Bria
- bria-video-eraser – Video → Video: Removes unwanted objects from videos.
4.2.2 Image Models
Bria
- bria-fibo – Text → Image: Generates brand-safe, commercially usable images from text prompts.
- bria-genfill – Image + Mask → Image: Fills or replaces selected regions with context-aware content.
- bria-eraser – Image → Image: Removes unwanted objects or elements from an image cleanly.
Google (Gemini)
- gemini-3-pro-image-preview – Text → Image: High-quality image generation with strong prompt understanding and visual accuracy.
Seedream / SeedEdit
- seedream-4-0-250828 – Text → Image: Produces high-fidelity images with rich details and consistent styles.
- seedream-3-0-t2i-250415 – Text → Image: Balanced text-to-image generation for general use.
- seededit-3-0-i2i-250628 – Image → Image: Advanced image editing while preserving structure and key details.
Tongyi (Alibaba)
- Z-Image-Turbo – Text → Image: Ultra-fast and cost-efficient text-to-image generation for rapid iteration.
Reve
- reve-create-20250915 – Text → Image: Creative text-to-image generation with strong artistic styles.
- reve-edit-20250915 – Image → Image: Prompt-based image editing while preserving layout and subject.
- reve-edit-fast-20251030 – Image → Image: Faster image editing for quick adjustments.
- reve-remix-20250915 – Image → Image: Remixes an image into new styles or variations.
- reve-remix-fast-20251030 – Image → Image: Fast image remixing with lower latency.
4.2.3 Audio Models
Note: When uploading audio to GMI, only WAV and MP3 file formats are supported.
Minimax Audio
- minimax-audio-voice-clone-speech-2.6-turbo – Text + Voice Sample → Audio: Fast voice cloning and speech generation with natural tone.
- minimax-audio-voice-clone-speech-2.6-hd – Text + Voice Sample → Audio: High-fidelity voice cloning with improved clarity and realism.
- minimax-tts-speech-2.6-turbo – Text → Audio: Low-latency text-to-speech for real-time or high-throughput use.
- minimax-tts-speech-2.6-hd – Text → Audio: High-quality text-to-speech with richer voice detail.
- minimax-tts-speech-2.5-turbo-preview – Text → Audio: Preview version optimized for fast speech generation.
- minimax-tts-speech-2.5-hd-preview – Text → Audio: Preview high-definition text-to-speech generation.
- minimax-tts-speech-02-turbo – Text → Audio: Efficient text-to-speech with balanced speed and quality.
- minimax-tts-speech-02-hd – Text → Audio: Enhanced speech quality with clearer pronunciation.
- minimax-tts-speech-01-turbo – Text → Audio: Earlier turbo TTS model optimized for speed.
- minimax-tts-speech-01-hd – Text → Audio: Earlier HD TTS model focused on audio clarity.
ElevenLabs
- elevenlabs-tts-v3 – Text → Audio: High-quality text-to-speech with expressive and natural voices.
- elevenlabs-tts-multilingual-v2 – Text → Audio: Multilingual text-to-speech supporting multiple languages and accents.
Step Audio
- Step-Audio-EditX – Audio + Voice Reference → Audio: Voice cloning and audio editing for modifying or recreating spoken content.
ComfyUI Nodes
- Native ComfyUI nodes sourced directly from the official ComfyUI repository
- Standard preprocessing, conditioning, and utility nodes commonly used in Comfy workflows
When to Use GMI Official API Nodes vs ComfyUI Nodes
- GMI Official API Nodes: Fully managed inference, simplified inputs/outputs, optimized performance
- ComfyUI Nodes: Fine-grained control, custom preprocessing, advanced workflow logic
4.3 Canvas Basics
- Add Node: Right‑click → Add Node / Find a node in node
- Delete Node: Select the node and click the trash symbol
- Drag & Arrange: Move nodes freely on the canvas
- Connect Nodes: Drag from an output of one node to an input or another; the data type must match for the connection to be valid
- Run: Execute the workflow
- Save: Save workflow changes
