
Reduced manual data entry and billing errors
Faster challan-to-invoice reconciliation
Real-time visibility into due bills
ZU Mechanical Works was managing critical billing and logistics workflows across disconnected channels: scanned documents, spreadsheet tracking, ad-hoc quotation drafts, and WhatsApp message threads. Teams were repeatedly retyping document fields, manually reconciling challans against invoices, and relying on inbox-based communication tracking that created response delays and audit gaps.
Neura engineered a Next.js 15 application with server-side API routes that orchestrate document ingestion, AI extraction, data normalization, and persistent storage in Supabase. Uploaded files are sent through Groq-powered multimodal extraction, normalized into a consistent schema, and stored with metadata and source URLs for traceability. This creates a reliable data foundation for downstream invoicing and reporting.
Users upload document images and the system extracts document number, dates, and item-level fields using Groq multimodal models. A normalization layer enforces a stable schema and reduces noisy OCR/LLM outputs before persistence.
Extracted documents, quotations, contacts, and WhatsApp events are stored in Supabase with queryable API routes. Teams can retrieve historical records quickly with filters for date ranges, identifiers, and business entities.
Dedicated UI modules allow teams to generate and preview challans, invoices, and quotations with structured line items. Accounts Receivable tracking provides visibility into outstanding balances, payment status, and aging summaries.
Webhook-driven WhatsApp ingestion centralizes inbound communication by contact and supports read-state updates. Built-in deduplication and metadata enrichment improve message quality and operational reliability.
A command-center dashboard aggregates invoice activity, receivables status, and monthly billing trends for quick operational awareness. Teams get immediate visibility into due payments, recent interactions, and execution pace.
PDF and image previews are supported. All documents and metadata are securely stored using Supabase Storage and Postgres.