VakeelOS - Legal Practice Operating System

The problem

Solo advocates and small Indian law firms still coordinate matters across paper causelists, messaging apps, spreadsheets, and disconnected document folders. Generic practice-management software misses court-specific workflows; generic AI adds a second risk by producing confident text without a reliable link back to the advocate's own case material.

The challenge

The product had to unify daily legal operations without flattening the way Indian advocates actually work. That meant CNR-linked matters, hearings, causelists, drafting, documents, fee collection, and phone-first access - while keeping primary data in Azure Central India and making non-India AI fallbacks an explicit opt-in. The interface and the underlying service boundaries needed to tell the same story: a dense professional system that still feels legible on first use.

Architecture & System Design

VakeelOS is a Next.js and TypeScript application organised around thin route adapters and domain modules. tRPC exposes typed application boundaries; Prisma and PostgreSQL with pgvector store operational and retrieval data; Azure Blob Storage handles documents; Inngest runs background jobs; and Redis supports cache and coordination paths. Azure OpenAI is the primary AI provider, while browser-side PaddleOCR keeps initial document extraction on-device. Production runs on Azure Container Apps with Key Vault, Log Analytics, Sentry, and fail-closed release checks.

Code Walkthrough

Non-proprietary extract demonstrating the core integration pattern.

// Provider policy: residency-safe by default
export function resolveAiProvider(env: ProviderEnv) {
  if (env.AZURE_OPENAI_RESOURCE && env.AZURE_OPENAI_KEY) {
    return createAzureProvider({ region: "centralindia" });
  }

  if (env.ANTHROPIC_API_KEY && env.AI_ALLOW_NON_INDIA_FALLBACK) {
    return createExplicitFallbackProvider();
  }

  throw new Error("No policy-compliant AI provider configured");
}

Results

The platform now has a production deployment path on Azure Container Apps, a modular route-and-service architecture, on-device OCR, grounded document retrieval, and explicit residency controls around AI providers. The same court-native design language carries through the product dashboard, mobile flows, brand identity, and responsive marketing site, so prospective users can understand the product before signing in.

Live Demos