🌿 Bioledger: I Built an AI That Wants Me to Touch Grass

Mohit Sharma — Mon, 12 Oct 2026 05:09:17 +0000

🌿 Bioledger: I Built an AI That Wants Me to Touch Grass

What if AI didn’t want more of your screen time? What if it helped you put your phone down and notice the world around you?

We have AI that writes emails, generates code, summarizes meetings, and answers questions we could probably have Googled.

So I tried something a little different.

I built Bioledger — a nature field notebook powered by open-source AI, designed to help you notice, record, and learn from the little things happening outside.

A bird on a wire. A strange-looking cloud. A leaf you’ve walked past a hundred times without really looking at it.

Nature has been doing its thing all along. Maybe we just need a better way to pay attention. 🍃

🐦 Meet Bioledger

🔗 Try it: https://bioledger.onrender.com

💻 Source code: https://github.com/MohitSharma2003/bioledger

Bioledger is a digital field notebook for curious people, nature lovers, beginner birdwatchers, and anyone who has ever looked at something outside and thought:

“Wait… what exactly am I looking at?”

Here’s what it can do.

📓 1. Log a Sighting

Spot something interesting? Record your observation and let AI help turn your notes into a more structured entry.

You don't need to be a professional naturalist. Sometimes an observation starts with something as simple as:

“There was a cloud this morning that looked like heaven.”

Not every observation needs a scientific name. Curiosity is a perfectly good place to start.

🔎 2. Build Your Own Nature Ledger

Your observations become a personal collection of sightings instead of disappearing into the camera roll or the depths of your notes app.

Revisit what you've seen, explore your entries, and slowly build a record of the natural world around you.

💬 3. Ask Your Journal

Ever wondered whether you recorded a particular bird before?

Ask your journal questions about your observations and explore what you've documented.

Instead of treating every interaction as a fresh conversation with a generic chatbot, Bioledger makes your own field notes the starting point.

📊 4. See Your Observations Come Together

Your ledger gives you a broader view of your activity, helping you explore your recorded sightings and notice patterns in your observations.

🌱 5. Explore What's In Season

Bioledger also includes an In Season section to make exploring nature feel more like an ongoing journey than a one-time checklist.

The goal isn't to turn a walk into another productivity metric. It's to give curiosity somewhere to grow.

🤖 The AI Behind the Leaves

For my local setup, I'm using Gemma 3 4B through Ollama.

That means I can run an open-weight language model on my own machine rather than relying entirely on a hosted AI API for every interaction.

Why does that matter?

More control: I can experiment with the model and how it behaves.
Local inference: In my local setup, AI processing can happen on my own machine.
Privacy potential: Keeping observations and inference local can reduce the need to send personal field notes to an external AI provider.
Less dependence on closed APIs: I can build around an open-weight model instead of tying the entire AI experience to one proprietary service.

One important distinction: running a model locally doesn't automatically make the entire application offline. The deployed demo and my local AI setup are different environments, and fully offline operation is something the architecture needs to support end to end.

I'm treating that as an engineering challenge, not a claim I'm going to pretend is already solved.

🌍 Why Open-Source AI?

For me, the interesting part isn't simply that a model can answer questions.

It's that I can choose how to run it, inspect the surrounding code, experiment with alternatives, and build an experience around my own goals.

With Bioledger, that goal is deliberately small:

Use AI to make the real world more interesting — not to replace it.

I don't need a model to tell me to spend three more hours chatting about a tree. I want it to help me record what I noticed, learn something useful, and get back to looking at the tree.

Open models make that kind of experimentation more accessible. I can build, test, and change the system without making a closed API the centre of everything.

And honestly, if AI can help someone become more curious about the patch of nature outside their home, that's a pretty good use of the technology.

🛠️ How I Built It

The project brings together:

Frontend: React, TypeScript, and Vite
Backend: Python and FastAPI
AI: Gemma 3 4B with Ollama in my local setup
Deployment: Render

The project is still evolving. I'm working toward a stronger offline-first experience, more capable observation workflows, and a setup that makes the boundary between local processing and the deployed application clearer.

Building this has also reminded me that adding AI is often the easy part. Making the whole experience useful, honest about its limitations, and pleasant to use is the real work.

🎬 See Bioledger in Action

I've included a short walkthrough showing the interface and its main features.

[Watch the demo video attached to this post]

You can also explore the live application and browse the code yourself:

🌿 Live demo
💻 GitHub repository

🚧 What's Next?

There's plenty more I'd love to build:

Better offline storage and synchronisation
More capable photo-based observation workflows
Easier model selection and local inference
A richer field notebook experience for people exploring their local biodiversity

But I don't want Bioledger to become another app that demands your attention all day.

If it helps you notice one new bird, recognise an unfamiliar plant, or simply spend a few more minutes looking at the sky, it's doing something right.

🌿 One Last Thing

We spend a lot of time asking what AI can do for us online.

For this project, I wanted to explore a different question:

Can AI help us pay more attention to the world when we're offline?

That's what I'm exploring with Bioledger.

If you try it, I'd love to hear what you think. What would you want in a nature journal like this — bird identification, plant observations, seasonal changes, or something else entirely?

And, most importantly, what's the last interesting thing you noticed outside? 👀

Built for the Hacktoberfest DEV Challenge — Week 1: Touch Grass.

#hf26challenge #ai #opensource

DEV Community: Mohit Sharma

🌿 Bioledger: I Built an AI That Wants Me to Touch Grass

🌿 Bioledger: I Built an AI That Wants Me to Touch Grass

🐦 Meet Bioledger

📓 1. Log a Sighting

🔎 2. Build Your Own Nature Ledger

💬 3. Ask Your Journal

📊 4. See Your Observations Come Together

🌱 5. Explore What's In Season

🤖 The AI Behind the Leaves

🌍 Why Open-Source AI?

🛠️ How I Built It

🎬 See Bioledger in Action

🚧 What's Next?

🌿 One Last Thing