KhataGO
WhatsApp-first accounting platform. Webhook → Gemini function-calling → idempotent ledger write. Handles transactions, receivables, and reconciliation entirely through WhatsApp chat + receipt photos.
Performance & Impact
< 1ms
Ingress dedup
Redis-backed message-id hash lookup drops duplicate Meta webhooks before any LLM call
8
Tool calls
Gemini function-calling surface: transactions, receivables, payments, ledgers, reminders
3
Languages
Full i18n across English, Hindi, and Gujarati for both UI and bot responses
The Problem
Small business owners in India struggle with complex accounting software. They often rely on manual notebooks (Khatas), which lead to data loss, calculation errors, and delays in GST compliance.
The Solution
A 'zero-learning-curve' platform that works where the user already is: WhatsApp. By combining the simplicity of chat with the power of AI, KhataGO makes business accounting as easy as sending a message.
System Architecture
KhataGO is a WhatsApp-first bookkeeping platform for Indian MSMEs. The backend is a reconciliation pipeline: Meta webhooks arrive with at-least-once delivery, a Redis-backed idempotency guard drops duplicates, Gemini 2.0 Flash with function-calling parses both text ('Sold 500 to Ram') and receipt images (OCR → structured JSON with merchant/amount/date/line items) into 8 tool calls — create_transaction, create_receivable, record_payment_received, get_party_ledger, send_payment_reminder, and others — each mapped to a Prisma write. Results flow back to the user over WhatsApp; a CA portal exports Tally-ready XML vouchers for month-end close.
Core Engineering Achievements:
System Architecture
Ingress
AI / Reconciliation
Ledger
Interface
"A reconciliation pipeline dressed up as a WhatsApp bot: events in, idempotency-guarded tool-calls transform them into ledger writes, export back to accountants' formats. The interesting engineering is upstream of the UI — dedup, function-calling determinism, double-entry prevention."
Visual Showcase
A high-fidelity walkthrough of the core interfaces and user experiences, designed with modern aesthetics.
Business Ledger Dashboard
A responsive, single-page summary tracking daily, weekly, and monthly incoming/outgoing financial trends.
Transaction Activity Feed
A chronological list of all categorized invoices, expenses, and payments with immediate export and edit capabilities.
Monthly Revenue & GST Portal
An integrated report generation interface allowing CA firms to download precise GST-ready and Tally XML statements.
The Engineering Challenge
Meta's WhatsApp Cloud API sends duplicate webhook events. The deduplication layer hashes the incoming message ID and checks a Redis TTL window before processing — duplicate events are dropped in under 1ms. The Gemini AI OCR pipeline downloads the WhatsApp image URL, sends it to Gemini Vision, and maps the extracted JSON (merchant, amount, date, line items) to a ledger transaction. Tally XML export is non-trivial: Tally's voucher schema requires specific date formatting, ledger name lookups, and GST field structure — wrong XML silently fails to import. Bull queues decouple all three operations so the WhatsApp bot responds immediately while processing happens in the background.
User Journey
Onboarding
Business owners register via the web portal and connect their WhatsApp Business API.
Master Data Setup
Configure customer/vendor records and categories in the web dashboard.
Transaction Logging
Send text messages or photos of bills directly to the KhataGO WhatsApp bot.
AI Processing OCR
The AI agent processes receipts via OCR and creates draft transactions automatically.
Reporting & Export
Review monthly GST summaries and export Tally-compatible XML vouchers for accountants.
Interested in the full engineering breakdown?
I'm always open to discussing technical implementations, from state management strategies to infrastructure scaling.