GenAI for Easy Read: Open-Source Multimodal Solution

Convert documents into accessible Easy Read formats with simplified text and supported symbols.

Step 1: Input Document or Text

Supports .txt, .md, .docx and .pdf — the text is loaded into the box on the right.

LLM
'hf' runs Llama via Hugging Face Inference (needs HF_TOKEN). 'llama'/'bedrock' use AWS; 'gemini' uses a Google key.
Translation
'nllb' loads a 1.3B model on first use — the only option for Luo and KSL gloss.
Translate into

Step 2: Review and Edit Easy Read

No sentences yet — add text above and press Simplify.

Step 3: Symbols and Icons

Symbol set

Image generation runs the LoRA model on the CPU Space — expect a few minutes per missing icon.

Read Aloud

Voice
0.5 1.5

Step 4: Export to Word