Handwriting Intelligence
20-factor analysis and personalised practice across five Indic scripts, live today. The analyser engine is open source.
Live Who buysSchoolsParentsCoaches Explore handwriting intelligenceOne pen · Six AI products · One motion dataset
For 100 years, pens recorded ink.
Now they record intelligence.
Every stroke the pen records becomes AI-ready motion data.
- Ordinary paper
- No camera
- No special notebook
- Works offline
- AI powered
- Patent protected
- Patent No. 584433
- Made in India
- DPIIT Startup
- RTIH Incubated
- DeepTech Innovation Award
Billions of pen strokes are written every day. Almost none of them become digital intelligence. Why that matters →
208 HzMotion capture
16-axisDual-IMU sensing
5 scriptsIndic + English analysis
No. 584433Granted Indian patent
The bridge · tradition, connected to digital
Handwriting is the brain’s daily calibration.
Your brain wasn’t designed to tap. Watch one letter being written. Every stroke runs a full learning loop.
Learning reinforced.
One letter, a complete learning cycle. Now imagine a whole page.
Typing
the same motion, repeated
vs
Handwriting
a new movement every letter
Digital devices make answers faster. Handwriting makes brains stronger.
- Better focus
- Better memory
- Better fine motor control
- Better learning
- Better handwriting
One page.
500 letters.
Thousands of brain loops.
That’s where Vahini begins.
Many products, one motion model
One signal. Six products today. Hundreds tomorrow.
Every product here runs on the same proprietary handwriting-motion dataset and one model. Handwriting analysis is the live proof; each new product both ships value and feeds more motion data back into the moat, so the platform compounds instead of being capped by any single product.
Digital Notes
Write on any paper, screen-free, and watch it digitise in real time.
In dev Who buysStudentsProfessionalsEnterprises Explore digital notesLearning Analytics
Classroom-level insight into how students write, for better outcomes.
Pilot Who buysSchoolsDistrictsGovernments Explore learning analyticsMotor Assessment
Early motor screening and intervention, studied with clinicians.
Research Who buysTherapistsHospitalsClinics Explore motor assessmentSignature Intelligence
Verify how a signature was written, not just how it looks, from the motion itself.
Next Who buysBanksInsuranceLegal Explore signature intelligenceDeveloper SDK
Motion intelligence as an API, build the future on Vahini for partners and researchers.
Future Who buysDevelopersPartnersAI labs Explore the API planOne pen · one signal · many markets
The pen is the core. Each market is the same signal, put to work.
Vahini came first: screen-free, paper-based, motion-aware. The 20-factor analyser exists to back the pen with a product people can use today. From there, the same motion signal reaches market after market, no new hardware, no new data.
VahiniThe core
A dual-IMU sensor ballpoint that writes on ordinary paper and records the motion of every stroke at 208 Hz, screen-free.
The source every market below draws from · Patent No. 584433HandwritingLive today
20-factor analysis & personalised practice across five Indic scripts, the product that backs the pen.
Coaches, parents, students & schoolsDigital notesIn build
Class notes written on paper, reconstructed from the IMU signal into a digital page. No tablet, no app open, no screen-time.
Students, schools & anyone who thinks better on paper · our founding ideaAssistive & screeningResearch
Support for writing difficulty and the early motor signals behind it, studied with clinicians, never a standalone diagnosis.
Special schools, occupational therapists, accessibility & research partnersThe datasetThe moat
India’s consented, multi-script handwriting-motion data, a defensible asset and a future licensing line.
What makes every market above hard to copyMarkets shown to illustrate scope; size each with primary research before external use.
Traction & recognition
Award-winning, patented, and already live.
Granted IP, government recognition and a live product, plus where we've shown Vahini. Newest first, updated as we go.
Latest updates
Where the work is going.
Notes from the research frontier and moments from the road, as we build the pen, the dataset and the science behind them.
Reconstructing handwriting from motion alone
A pen only touches the paper part of the time. Telling writing from hovering is the heart of motion-based handwriting.
Read the post
DeepTech Innovation Award winner
Recognised for the dual-IMU pen and the 20-factor analyser, alongside a granted patent and DPIIT recognition.
See the recognition
Showing the pen, hand to hand
From T-Hub to IIT and MSME forums, demoing how motion on ordinary paper becomes an actionable handwriting report.
See where we’ve beenFrom the road
What people said at our events.
Notes left in our book by educators, founders and students who tried Vahini in person.
Proven first in handwriting · Live today
Proof the signal becomes real insight.
Our handwriting solution is live now. Upload any handwritten page and get a 20-factor analysis with personalised exercises, across five Indic scripts. It's the working proof of everything the platform promises.
The asset nobody else holds
India’s consented multi-script handwriting dataset.
Every analysed page and every pen-stroke, captured with explicit consent, builds a corpus that doesn’t exist anywhere else: real handwriting motion across five Indic scripts and many ages. A camera company can scrape finished pages; only Vahini owns the motion behind them.
Consent-firstCollected only with explicit consent, never quietly scraped.
Motion, not contentWe store the signal, pressure control, pen control, writing flow, never the words written.
DPDP-aligned, India-processedPrivacy by design under the DPDP Act 2023, a sales advantage, not just compliance.
The corpus
Five scripts, two signal layers, one compounding moat.
Englishहिन्दीతెలుగుதமிழ்ಕನ್ನಡ
The technology
Built on real hardware, on real paper.
The pen reads motion, not the page, so it works on whatever paper you already use.
Dual-IMU sensing
Two inertial units (one 9-axis with magnetometer, one 6-axis) plus a tip force sensor stream 16 axes at 208 Hz, Kalman-filtered.
Works on any paper
No dot-pattern, no tablet, no special notebook. The signal comes from the pen, not the page.
Five-script modelling
Motion models tuned across five Indic scripts, the basis of a consent-first motion dataset.
Patented
Indian Patent No. 584433, DPIIT-recognised.
One pen. Regular paper. Endless signal.
Try the analyser that's live today, or help build the platform behind it.