🚀 Quick Start Guide
1. Visual Setup (The "Magic Box"):
  - Drag the Nocoddo AI element onto your page.
  - This element is visual: it displays the "Thinking..." spinner, the download progress bar, and the final result text automatically.
  - Customize the Background, Border, and Font in the property editor to match your app's design.

2. Running the AI:
  - Add a Button (e.g., "Analyze").
  - Start a Workflow: Element Actions > Run AI.
  - Main Text / Context: Pass the text you want to analyze (e.g., `Input A's value`).
  - Task: Select the AI operation (see Documentation for details).

⚠️ IMPORTANT: First Run Behavior
Because this AI runs on the user's device, the first time a user runs a specific task, the browser must download the AI Model (between 40MB and 250MB).
  - Action: A progress bar will appear inside the element showing the download status.
  - Advice: Warn your users that the first analysis might take 10-60 seconds. Subsequent runs will be instant (cached).

3. Batch Processing (Database Items):
  - Since this runs Client-Side, you cannot use Backend Workflows.
  - To process a list of database items (e.g., summarize 500 tickets), create an "Admin Worker Page".
  - Create a recursive workflow on that page: Load Item -> Run AI -> Save to DB -> Trigger Next Item.
  - Keep the tab open while it processes the queue for free.

4. Storage Management:
  - Use the Clear AI Cache action to free up storage space on the user's device if needed.



Where to find compatible models?

You cannot use just any model (like standard GPT or Llama). You need models optimized for the web (ONNX format) supported by Transformers.js.

1. Go to Hugging Face Hub: https://huggingface.co/models

2. Filter for Transformers.js: Search for models with the tag transformers.js or look at the Xenova collection (highly recommended).

3. Copy the Model ID: Copy the text that looks like User/ModelName.

Example: Xenova/distilbert-base-uncased


Recommended Custom Models by Task

If you want to change the default behavior, here are the best tested models for each task supported by this plugin.

  - Sentiment Analysis: Xenova/bert-base-multilingual-uncased-sentiment
  - Summarization: Xenova/bart-large-cnn
  - Text Generation: Xenova/flan-t5-base
  - Zero-Shot Classif.: Xenova/bart-large-mnli
  - Q&A English: Xenova/roberta-base-squad2
  - Q&A Multilingue: Xenova/xlm-roberta-base-squad2
  - Feature Extraction: Xenova/paraphrase-multilingual-MiniLM-L12-v2


⚠️ Important Performance Warning

Size Matters: Since these models run in the browser, the user must download them once.

Small Models (<100MB): Fast load, good for mobile.

Large Models (>500MB): High accuracy, but may take 10-30 seconds to initialize on the first run and might crash on older mobile devices.

Always test the Custom Model ID on your target device before deploying.