Models & Providers

AI models and local vs cloud options.

This guide explains AI models in NodeTool and how to choose between local and cloud execution.

What Are AI Models?

An AI model is a trained program for a specific task:

Model Type Function Example
Language Model (LLM) Text generation Stories, Q&A, summarization
Image Model Image generation/editing Artwork, photos, variations
Speech Model Speech↔text conversion Transcription, text-to-speech
Vision Model Image understanding Photo descriptions, OCR

Models come pre-trained. Select one for your task.

Local vs. Cloud

NodeTool runs AI models locally or through cloud APIs.

Local Models

Pros:

  • 🔒 Private – Data stays local
  • 💰 Free – No usage costs
  • 📶 Offline – Works without internet

Cons:

  • 💾 Requires space – 4-15 GB per model
  • ⚡ Needs hardware – Faster with GPU
  • ⏳ Initial download – One-time setup

Cloud Models

Pros:

  • 🚀 Fast – No downloads
  • 💻 Any hardware – Works on older machines
  • 🆕 Latest models – Access newest capabilities

Cons:

  • 💵 Usage costs – Pay per task
  • 🌐 Requires internet
  • 📤 Data sent externally

Combine local and cloud:

  • Speech recognition – local for privacy
  • Image generation – cloud for quality
  • Document processing – local for confidential files

State-of-the-Art Cloud Models for Creative Workflows

NodeTool provides access to the latest high-quality generative AI models through cloud providers. These models represent the cutting edge in video, image, and audio generation:

Top Video Generation Models

Model Provider Capabilities Key Features
OpenAI Sora 2 Pro OpenAI T2V/I2V up to 15s Realistic motion, refined physics, synchronized native audio, 1080p output
Google Veo 3.1 Google T2V/I2V with references Upgraded realistic motion, extended clip length, multi-image references, native 1080p with synced audio
xAI Grok Imagine xAI T2V/I2V/T2I Multimodal text/image to short video with coherent motion and synchronized audio; also text-to-image
Alibaba Wan 2.6 Alibaba Multi-shot T2V/I2V Affordable 1080p with stable characters and native audio; reference-guided generation
MiniMax Hailuo 2.3 MiniMax High-fidelity T2V/I2V Expressive characters, complex motion and lighting effects
Kling 2.6 Kling T2V/I2V with audio Text/image to synchronized video with speech, ambient sound, and effects; strong audio-visual coherence

Top Image Generation Models

Model Provider Capabilities Key Features
Black Forest Labs FLUX.2 BFL T2I with control Photoreal images, multi-reference consistency, accurate text rendering, flexible control
Google Nano Banana Pro Google High-res T2I Sharper 2K output, intelligent 4K scaling, improved text rendering, enhanced character consistency

Using These Models

Access these models through NodeTool’s generic nodes:

  1. For Video: Use nodetool.video.TextToVideo or nodetool.video.ImageToVideo
  2. For Images: Use nodetool.image.TextToImage
  3. Select Provider: Click the model dropdown in the node properties
  4. Configure API: Add provider API keys in Settings → Providers

Direct NodeTool API Key Support:

  • OpenAI Sora 2 Pro: OPENAI_API_KEY
  • Google Veo 3.1: GEMINI_API_KEY
  • MiniMax Hailuo 2.3: MINIMAX_API_KEY

Access via kie.ai (recommended for other models):

  • xAI Grok Imagine, Alibaba Wan 2.6, Kling 2.6, Black Forest Labs FLUX.2, Google Nano Banana Pro
  • Configure using KIE_API_KEY in Settings → Providers

Cost Considerations: Cloud models typically charge per generation. Check each provider’s pricing before extensive use. Local models are free after download but require capable hardware.

Alternative Access: Many of these models are available through kie.ai, an AI provider aggregator that often offers competitive or lower pricing compared to upstream providers. For models without direct NodeTool API key support (xAI, Alibaba, Kling), kie.ai is the recommended access method.

Getting Started

  1. Open Models → Model Manager in NodeTool
  2. Install these starter models:
    • GPT-OSS (~4 GB) – Text generation and chat
    • Flux (~12 GB) – High-quality image generation
  3. Wait for downloads to complete
  4. Run templates – they’ll work offline!

Option 2: Start with Cloud Providers

  1. Get an API key from a provider:
  2. In NodeTool, go to Settings → Providers
  3. Paste your API key
  4. Select the provider when using AI nodes

Understanding Model Names

Model names can look confusing, but they follow patterns:

Name Part Meaning Example
Base name The model family GPT, Llama, Flux
Number Version/size GPT-4, Llama-3
Size indicator Capability level mini, small, large
Quantization Compression level Q4, Q8 (lower = smaller file)

Example: llama-3-8b-instruct-Q4 means:

  • Llama version 3
  • 8 billion parameters (medium size)
  • Instruction-tuned (follows directions well)
  • Q4 quantization (compressed to save space)

Don’t memorize this – NodeTool’s Model Manager shows compatible models for each task.

Detailed Guides

General

Local AI

Cloud AI

Advanced

Quick Reference: Common Tasks

“I want to generate text”

  • Local: Install GPT-OSS or Llama model
  • Cloud: Use OpenAI GPT-4 or Anthropic Claude

“I want to create images”

  • Local: Install Flux or Stable Diffusion
  • Cloud: Use OpenAI DALL-E or Fal.ai

“I want to transcribe audio”

  • Local: Install Whisper (recommended for privacy)
  • Cloud: Use OpenAI Whisper API

“I want to understand images”

  • Local: Install a Vision model (Llava, Qwen-VL)
  • Cloud: Use GPT-4 Vision or Claude with images

FAQ

Q: Do I need a powerful computer?
A: For local models, a GPU helps but isn’t required. Cloud providers work on any computer.

Q: How much do cloud models cost?
A: Typically $0.001-0.03 per task. Most providers offer free credits.

Q: Can I switch models later?
A: Yes. Use the Model button on any AI node to change models without rebuilding the workflow.

Q: Local or cloud?
A: Depends on your needs. Try both.