MoWoB
Start a handoff

Products, not portfolio pieces.

Three products we designed, built and run. Two more we built for clients, where we can describe the engineering but cannot show you the product. Every screenshot here is a capture of the live thing, never a mockup.

The Crowta home page: the headline “Your website. Built and run by chat.” beside a phone showing a WhatsApp conversation with the Crowta assistant.
SaaS Small business India

Crowta

A website you get by chatting. Owner-led Indian businesses describe what they need over WhatsApp, Telegram or Slack, by voice or text, in English, Hindi, Tamil, Bengali, Gujarati or Malayalam, and have a live site in about thirty minutes, updated for as long as they keep the conversation going. It sells into cafés and catering, clinics, fitness studios, and legal and finance practices, from ₹14,999 a year.

crowta.com ↗

Built by MoWoB

What we built
The whole product. The chat that reads a plain-language brief over WhatsApp, the generator that turns it into a working site, and the language handling that lets an owner write in Hindi, Tamil, Bengali, Gujarati or Malayalam and get the same result.
Status
Live at crowta.com. A site goes up in about thirty minutes and keeps updating from the same conversation, so the build and the maintenance are the same act.
The Voltavic home page: the headline “Leads. Projects. Sites. Finance. Team. One Platform.” beside a panel showing an electricity bill being turned into a bill of materials.
SaaS Solar EPC India

Voltavic

An operations platform for Indian solar EPC companies, covering the whole project lifecycle in one place. A sales CRM turns an uploaded electricity bill into a bill of materials; a nine-stage pipeline tracks execution with GPS-verified photo evidence; sites are monitored live across inverter brands; and OPEX/PPA billing, serial-number inventory, contractor payments and DLT-compliant SMS and WhatsApp all run off the same system. From ₹4,999 a month.

voltavic.com ↗

Built by MoWoB

What we built
The whole platform. The CRM that turns an uploaded electricity bill into a bill of materials, the nine-stage pipeline that tracks execution against GPS-verified photos, live monitoring across inverter brands, and the billing, inventory and contractor payments that all run off the same records.
Status
Live at voltavic.com, running solar EPC projects from first quote to generation.
The onco.fit home page: the headline “One chat. A clear path through cancer.” beside a chat window where a patient says they were advised many expensive tests.
Healthcare Oncology Delhi NCR

onco.fit

Cancer navigation for patients in Delhi NCR, run over web chat and WhatsApp. It organises a patient's reports, sets them on one of ten evidence-based oncology pathways, puts the case in front of a multidisciplinary cancer board, and checks what a test or treatment will actually change before anyone pays for it, connecting patients to government schemes and financial support along the way. Navigation is free; a ₹2,500-a-year care plan adds doctor consultations. Built by docOPD Technologies.

Its own framing is the sharpest thing about it: "We are the map, not another destination." It doesn't compete with hospitals. It routes patients across public, charitable and private providers, and says plainly on every screen that it is never a replacement for your doctor.

sutra.virinchisoftware.com/oncofit ↗

Built by MoWoB

What we built
The whole platform. One chat, on web or WhatsApp, that tells a patient what to do next, with the pathway steps, the government schemes and the real costs behind it. The guidance follows an oncology board rather than the model's own opinion.
Status
Live and in use across Delhi NCR. Every screen says plainly that it is not a replacement for a doctor, and nothing we wrote here softens that.
Pipeline
Capture under controlled lighting → face localisation and crop → facial point detection → skin masking → per-condition scoring
Scores
Scarring · Acne and spots · Texture · Dark circles · Wrinkles. Each returned as a percentage, not a diagnosis.
Runs on
Core ML, TensorFlow Lite, OpenCV, running on device, iOS and Android
Who reads the result
A dermatologist. The software scores what it can see and hands that to a clinician, who decides what it means and what to do about it. It does not diagnose, and it does not recommend treatment.
Healthcare Dermatology Computer vision

Clinical Face Analysis

An AI and machine learning face analyser for dermatology, running on a smartphone. It finds and crops the face, detects facial landmarks, masks the skin region, then scores spots and scars, texture, dark circles and wrinkles, so a patient's skin can be tracked over time and a dermatologist can prescribe against something measured rather than remembered.

Built by MoWoB, under contract

What we built
The analyser end to end. The capture step that holds lighting steady enough for two scans to be worth comparing, the vision pipeline that isolates the face and then the skin within it, and the scoring that turns what it sees into a figure per condition.
Why it runs on the phone
On the phone, not on a server. The models ship inside the app, so the photographs never leave the device. That matters more here than it usually would, because the image in question is somebody's face.
Why there is nothing to click
The contract. We can describe how it was built, and that is what is above. We cannot name the client, show the interface, or link you to a demo. If that matters for your decision, ask us on a call and we will tell you what we are able to.
Interfaces
Text and voice
Built on
A large language model reads what the person actually wrote and answers, rather than matching an intent from a fixed list
Deployed to
Web, iOS, Android, Windows, macOS, Linux
AI/ML NLP Automation

AI/ML Chatbots

Text and voice agents that talk like a person, run on a large language model briefed on the business, and deploy across web and native platforms, to take the routine half of a support or sales conversation off a team's plate. Crowta and onco.fit, above, are both chat-first, so the same ground is well covered.

Built by MoWoB

The problem
A small store cannot put somebody on WhatsApp and email all day. The volume does not justify a full-time person, but the questions still arrive, and a question nobody answers is usually an order nobody places.
What it handles
Product questions, pricing, discounts and current offers. It will take an order, and it will report on one already placed. Inside the rules the store sets, not whatever the model would say if left to itself.
Why there is nothing to click
Each one lives inside a client's own store and their own WhatsApp number, so there is no public bot to message. We can walk you through a real conversation on a call.

Your product could be next.

Send us the repo. We'll tell you what's wrong with it before you commit to anything.

Start a handoff