README.md

metadata

title: Comfortool
emoji: 🐨
colorFrom: blue
colorTo: green
sdk: gradio
app_file: app.py
pinned: false
tags:
  - building-mcp-track-consumer
  - building-mcp-track-creative

Comfortool

Autism-friendly social stories with calming illustrations

Comfortool is an MCP server and Gradio application designed to generate personalized social stories and calming illustrations for children with autism. The project leverages LLMs and generative AI to help children understand and navigate everyday scenarios in a supportive, accessible way.

Features

• Select a scenario and generate a simple, autism-friendly social story.
• Generate a calming illustration for each story.
• MCP tool integration (e.g., echo tool for testing).
• Extensible architecture for future tools.

Future Directions

Here are some ideas for future MCP tools and features to add more value:

• Personalized adaptation by age and aesthetics:
  Adapt both the story and illustration to the child's age, interests, and preferred visual style.

• Emotional feedback tool:
  Allow the child or adult to indicate how they feel about the story or situation, and adapt the content to provide personalized emotional support.

• Visual summary tool:
  Generate a simplified version of the story as pictograms or comics for children with reading difficulties.

• Q&A tool:
  Enable the child or adult to ask questions about the story or situation, and receive simple, supportive answers.

• Personalized routines tool:
  Create stories and guides for daily routines (getting up, going to school, etc.) tailored to the child's needs.

• Translation and cultural adaptation tool:
  Translate the story and adapt names, places, and customs to the child's language and culture.

• Gamification tool:
  Add small challenges, rewards, or games related to the story to increase motivation and learning.

• Progress tracking tool:
  Save stories, record reactions, and suggest new stories based on the child's progress.

Track

This project participates in Track 1: Building MCP
Category: Consumer MCP Servers
Tag: building-mcp-track-consumer

How to run locally

uv run app.py

How to deploy

Push your code to Hugging Face Spaces and configure your secrets (e.g., OPENAI_API_KEY).

---

Check out the Spaces configuration reference for more details.