DEV Community

Mithilesh Parmar
Mithilesh Parmar

Posted on

I left software to run a psyllium mill, so I shipped a pip package for ingredient buyers

I'm a software engineer by background. This year I started RM Psyllium, a psyllium husk mill in Sidhpur, Gujarat, the region where most of the world's Plantago ovata is grown.

The first thing I noticed on the trade side: buyers ask the same handful of questions on every inquiry. Which grade? What mesh? What should be on the COA? How do I ask for a quote that a supplier can actually price?

Engineers have a habit for repeated questions. We write a tool. So I wrote a tiny one.

pip install psyllium

psyllium grades      # purity grades and what each is used for
psyllium checklist   # what to ask a supplier for on every lot
psyllium quote       # what to include in a price request
psyllium certs       # certifications to look for
psyllium about       # who made this
Enter fullscreen mode Exit fullscreen mode

Zero dependencies, Python 3.7+, MIT licence. It's about 3 KB.

Why a CLI for an agricultural product?

1. The domain is under-documented. Psyllium husk (also sold as isabgol or ispaghula) goes into fiber supplements, gluten-free bread, plant-based meat binders and pharma laxatives. Yet the practical buying knowledge lives in email threads and in the heads of a few traders. Putting it in a versioned package makes it citable and correctable.

2. Grades are a lookup problem. Husk is sold from 95% to 99% purity, and powder by mesh size. Formulators mostly need a quick table, not a sales call.

3. Checklists beat memory. The checklist command prints what a buyer should request on every lot: purity, swell volume, moisture, ash, microbiology, heavy metals, pesticide residues and the lab that tested it. Missing one line on a COA is how a container gets held at customs.

What I learned shipping it

  • Packaging a 100-line script for PyPI still takes an afternoon: pyproject.toml, an entry point, a README that renders, and 2FA on the account.
  • Naming matters. psyllium was free, which says something about how little tooling exists in this space.
  • Plain text output is the right UI here. People paste it straight into a spec sheet or an email.

What's next

I'd like to add a landed-cost helper (FOB price plus freight, duty and insurance to a destination port) and a COA parser that flags out-of-spec values. If you work in food tech, supplements or procurement software and want to contribute, the package is on PyPI: pypi.org/project/psyllium

More about the mill: rmpsyllium.com

Top comments (0)