Introducing the EDS Coach: A Custom GPT Tailored for Those with Hypermobility Conditions

Chanelle Henry
3 min readFeb 18, 2024


In a world that’s rapidly embracing AI to solve an array of challenges, the intersection of technology and healthcare has never been more exciting — or promising. As someone deeply entrenched in the tech world and navigating the complexities of life with Ehlers-Danlos syndrome (EDS), I’ve wanted to harness the power of Generative Pre-trained Transformers (GPTs) for something profoundly personal and impactful: creating the EDS Coach, a custom GPT designed to support individuals with hypermobility conditions.

The Genesis of EDS Coach

Living with EDS or any hypermobility spectrum disorder (HSD) is akin to navigating a maze without a map. The conditions are complex, often misunderstood, and come with a wide range of symptoms that can be overwhelming. As a Senior UX Designer and Researcher transitioning into data science and prompt engineering, I’ve experienced firsthand the gaps in accessible, personalized support for those with EDS.

This realization sparked an idea: What if we could create a digital coach, powered by AI, specifically designed to offer guidance, information, and support to those with hypermobility conditions?

How EDS Coach Works

The EDS Coach is more than just a chatbot; it’s a comprehensive support system that leverages the sophisticated understanding capabilities of GPTs to provide personalized advice and information. By training the EDS Coach on a vast dataset that includes medical research, patient testimonials, and therapy insights specific to EDS and HSD, the tool can offer users:

  • Personalized Exercise Recommendations: Tailored suggestions to strengthen muscles and improve joint stability, taking into account the user’s specific pain points and limitations.
  • Dietary Guidance: Advice on nutrition that supports joint health and overall well-being, considering any comorbid conditions common among those with EDS.
  • Pain Management Strategies: Techniques and practices to manage chronic pain, from mindfulness exercises to understanding when to seek medical intervention.
  • Community Connection: Links to EDS support groups and forums, offering a space for shared experiences and advice.

Navigating Challenges with Empathy and Precision

One of the key challenges in creating the EDS Coach was ensuring that the advice provided was not only accurate but delivered with empathy and understanding. EDS and HSD affect everyone differently, making personalized, nuanced responses critical. The development process involved iterating on the model’s responses to ensure they were informative, supportive, and reflective of the diverse experiences of those within the EDS community.

Sample output with the Coach

The Road Ahead

The EDS Coach is a step toward a future where technology bridges the gap between medical knowledge and the day-to-day realities of living with chronic conditions. As we continue to refine and expand its capabilities, the potential to make a tangible difference in the lives of those with EDS and HSD grows.

This project is a testament to the power of combining personal experience with technological expertise. It’s about leveraging my background in UX design, data science, and prompt engineering to create something that not only informs but empowers and supports. As we look to the future, the EDS Coach stands as a beacon of how custom GPTs can be tailored to meet specific healthcare needs, offering hope, support, and guidance to those who navigate the complexities of hypermobility conditions every day.

In sharing this journey, I hope to inspire others in the tech and healthcare communities to think creatively about how AI can be harnessed to address unmet needs, making a real difference in the lives of individuals around the world. The intersection of technology and healthcare is ripe with opportunities for innovation, and with projects like the EDS Coach, we’re just beginning to scratch the surface.

If you have EDS, please check out this plugin, as I would love your feedback to be able to build and train this model even further. I’m excited for what’s to come.

Try it out here: