The Replicate node enables seamless integration with Replicate's API for running machine learning models in your workflows. This node allows you to leverage Replicate's extensive collection of pre-trained models directly within your Nodetool pipelines.
- Direct integration with Replicate's API
- Real-time model status monitoring
- Support for various model types and inputs
- Automatic error handling and status reporting
The Replicate node includes several specialized audio processing nodes:
- AudioSR: Upscale and enhance audio quality using AI super-resolution
- Supports configurable parameters like seed, steps, and guidance scale
- Ideal for improving audio fidelity and clarity
- Riffusion: Generate music using stable diffusion
- MusicGen: Create music from text prompts or existing melodies
- TortoiseTTS: Text-to-speech with voice cloning capabilities
- StyleTTS2: Advanced text-to-speech generation
- MMAudio: Add AI-generated sound to videos
- RealisticVoiceCloning: Create song covers with AI voice models
- Demucs: Professional audio source separation
- Separate tracks into vocals, drums, bass, and other instruments
- Multiple model options (htdemucs, mdx_q, etc.)
- Support for various output formats (MP3, WAV, FLAC)
- IncrediblyFastWhisper: Fast audio transcription and translation
- Support for 90+ languages
- Word-level and chunk-level timestamp options
- Optional speaker diarization
The Replicate node includes several specialized image processing nodes:
- CLIP Interrogator: Generate text prompts that match images
- Multiple modes: best, fast, classic, negative
- Support for different CLIP models
- Moondream2: Efficient vision language model for edge devices
- BLIP/BLIP2: Generate image captions and answer questions about images
- LLaVA-13b: Visual instruction model with GPT-4 level capabilities
- NSFW Detection: Detect inappropriate content in images
- CLIP Features: Extract features from images using CLIP models
- CodeFormer: Restore old photos and AI-generated faces
- Face upsampling and background enhancement
- Adjustable quality/fidelity balance
- Night Enhancement: Improve low-light images
- SUPIR: Photo-realistic image restoration
- Multiple versions (v0Q, v0F) for different quality levels
- Advanced color correction and enhancement
- MAXIM: Multi-purpose image processing
- Denoising, deblurring, deraining
- Indoor/outdoor dehazing
- Low-light enhancement
- Old Photos Restoration: Specialized restoration for vintage photos
- FaceToMany: Transform faces into different styles
- 3D, emoji, pixel art, video game, claymation, toy styles
- Adjustable strength and control parameters
- BecomeImage: Adapt faces to match reference images
- PhotoMaker: Create stylized portraits
- Multiple style presets (Disney, Digital Art, Comic Book, etc.)
- Support for multiple input images
- FaceToSticker: Convert faces into sticker-style images
- InstantID: Advanced face transformation
- Multiple variants (Photorealistic, Artistic)
- Extensive control over generation parameters
- Support for pose control and face enhancement
- Sign up for a Replicate account at https://replicate.com
- Get your API token from the Replicate dashboard
- Configure your API token in the Nodetool app:
REPLICATE_API_TOKEN
- Add a Replicate node to your workflow
- Select the desired model from Replicate's collection
- Configure the model inputs according to the model's requirements
- Connect the node to your workflow's input and output nodes
- Run your workflow
The node will automatically handle:
- API authentication
- Model inference requests
- Status monitoring
- Result processing
The Replicate node includes a status indicator that shows:
- 🟢 Online: Model is available and ready
- 🔴 Offline: Model is currently unavailable
The node includes built-in error handling for common issues:
- Invalid API tokens
- Model availability issues
- Input validation errors
- Network connectivity problems
Contributions are welcome! Please feel free to submit a Pull Request.
AGPL