Become a sponsor to Vahini Technologies
A pen that remembers how you wrote.
Vahini started with an exam a student could not write without someone else turning
up. We thought writing should not depend on that.
So we built a pen: an ordinary ballpoint with two motion sensors inside. Any paper,
no camera, no special notebook. It records every stroke at 208 Hz and we rebuild
what was written and how: speed, pressure, hesitation. Indian Patent No. 584433.
People keep writing to us with uses we had not planned: homework that stays on paper
but reaches a learning platform, proof a page was handwritten, case notes made
searchable, clinical findings that reach the record without retyping, notes a
patient with limited vision can read on their phone.
Project Akshara is our priority this year: the first handwriting-motion dataset
for Indian scripts, recorded with schools, consent-first, in Telugu, Tamil, Hindi,
Kannada and English. Everyone photographed the page; nobody recorded the hand. This
is what sponsorship pays for most directly: collection sessions, consent and
cleaning, and releasing what can be released. Read more.
Free to use today
- 20factor-analyser, AGPL-3.0,
live at vahinitech.com - imu2text, handwriting from motion alone,
Apache-2.0 - fine-motor-skills-research,
Apache-2.0 - bujo.js, MIT
On the bench, waiting behind the dataset
- 20factor-analyser-pro: compare two pieces of writing or one writer over months;
dysgraphia and dyslexia screening as research with specialists; predictions. - Clinical notes into the medical record without retyping.
- Notes a patient with limited vision can read on their own phone.
- imu2text in Telugu, Tamil, Kannada and Hindi.
Sponsors at $100 and above tell us which moves first. Think the order is wrong?
Say so: info@vahinitech.com, or an issue on any repo. Sponsors' issues are answered
first.
Our focus is software; the pen is the hardware that follows, and what we sell.
Sponsorship funds the side we give away. Bootstrapped, DPIIT-recognised, RTIH
incubated, DeepTech Innovation Award 2026. The pen is in pre-order.
First goal: ten monthly sponsors.
Meet the team
-
Mallikarjunarao Kosuri vkosuriMachine Learner, Developer and Tester
-
KVR Vishnu VISHNU0906Software engineering, machine learning and product security. Founding engineer, patent inventor and security researcher.
Featured work
-
vahinitech/imu2text
IMU2Text: A hybrid CNN+GNN pipeline for handwriting recognition and trajectory prediction using IMU data with state-of-the-art accuracy (71.6%).
Python 6 -
vahinitech/fine-motor-skills-research
Research, worksheets, and AI algorithms for Fine Motor Skills and early childhood handwriting development.
-
vahinitech/bujo.js
A JavaScript library for generating customizable bullet journal PDFs with templates for planners, trackers, and more.
JavaScript 1 -
vahinitech/docuwriter
AI-powered text formatter for Vahini AI Pen.
Python -
vahinitech/20factor-analyser
Vahini Computer Vision 20 Factor Handwriting Analysis
Python -
vahinitech/ledger-book
Small Business Ledger Book
Python
0% towards 10 monthly sponsors goal
Be the first to sponsor this goal!
$5 a month
SelectYou keep the free analyser free. Your name on the supporters list, the sponsor badge, and a monthly note on what shipped and what broke. Ten of you is our first goal.
$25 a month
SelectEverything in Supporter, plus early access to analyser and imu2text releases and a say in what we build next: roadmap issues first, every comment read. For people who build on the pen's motion data.
$100 a month
SelectEverything in Builder, plus research notes as we write them (including what fails), early access to Project Akshara releases, and a priority reply to technical questions. This tier pays for collection sessions in schools, and you help decide which bench project moves first.
$500 a month
Select🚀 Technology Partner
Everything in Research Backer, plus your logo on vahinitech.com, a quarterly call with the team, and first access to prototype pens. For labs, schools and companies who want a hand in where this goes.