BaseLine

Inspiration

One of our team members has a grandfather whose early signs of cognitive decline went unnoticed for too long. Not because nobody cared, but because there was nothing quietly watching. GP appointments are short and infrequent. Decline is slow and invisible. By the time it becomes obvious, you have already missed the window where early action matters most.

We built Baseline because that window should not be missed.

What We Built

Baseline is a Chrome sidebar extension that runs three short daily tests, a typing dynamics test, a voice steadiness test, and an eye reaction test. Each session takes about three minutes. The app builds a personal picture of what normal looks like for each user and flags when scores start to drift from that baseline over time.

Before a GP appointment, one click generates a clean clinical PDF report written by Manus AI from the user's full session history. Something concrete and readable that a doctor can actually use.

How We Built It

We used React, Vite, and Tailwind CSS for the extension itself, with chrome. storage.local keeping all data entirely on the user's device. The Web Audio API handles the voice test feature extraction without ever storing audio. ElevenLabs reads test instructions aloud, making the app accessible for elderly users. Z.ai's GLM model generates a plain English explanation after each test completes. jsPDF renders the final GP export.

Before writing any code, we used Manus AI to research whether our three tests had real clinical grounding. It browsed medical literature and NHS documentation and came back with a sourced report that directly shaped our test protocols, scoring thresholds, and the language used in the GP export.

Challenges

Getting the Web Audio API pitch extraction to work accurately in a sidebar context was harder than expected. Balancing clinical credibility with language simple enough for a 70 year old to understand without feeling scared took a lot of iteration. Building something that feels like a daily habit rather than a medical device was the design challenge running through everything we did.

What We Learned

That the gap between clinical research and everyday health tools is enormous, and that closing it does not require a hospital or a research budget. It requires consistency, a bit of signal processing, and something worth coming back to every day.

Built With

• chrome-extensions-api-(manifest-v3)
• chrome.storage.local
• elevenlabs-api
• javascript
• jspdf
• manus-ai
• react
• tailwind-css
• vite
• web-audio-api
• z.ai-glm-api