features index

Features

Active contributors: Saksham, Ravi

360 Ghar is a unified real estate platform that bundles six product modules and several cross-cutting capabilities into a single async backend. This section of the wiki covers the user-facing features, one page per module. Each page traces its claims back to source files under app/ and cross-links the underlying systems where relevant.

Modules

The platform is organised around these feature areas, each with its own REST surface under /api/v1 and (for several) dedicated MCP servers tool surfaces:

• Ghar Core — the property marketplace. Swipe-based discovery, PostGIS geospatial search, PostgreSQL full-text search, hybrid semantic search over pgvector, and the recent 3-tuple cursor pagination refactor for list endpoints. Covers properties, swipes, visits, and agents.
• 360 Stays — short-stay bookings for hotels, vacation rentals, and temporary accommodation. Covers availability checks, dynamic pricing, the booking lifecycle, and the deliberate overlapping-bookings business rule.
• Flatmates — flatmate and PG discovery with swipe-based matching, conversations, moderation, QnA, and SSE-driven real-time events.
• Property Management — the PM system for landlords and relationship managers: leases, rent collection, maintenance, documents, inspections, reports, tenants, applications, and RM assignments.
• 360 Virtual Tours — immersive 360° tour platform with AI-powered hotspot generation, floor plans, analytics, and custom branded domains.
• 360 Data Hub — real estate data aggregation. Twenty-six scraper modules covering bank auctions, RERA projects, circle rates, gazette notifications, jamabandi, zoning, and neighbourhood scores, scheduled by a single APScheduler instance.

Cross-cutting capabilities

These features span multiple modules and are documented separately:

• MCP servers — two Model Context Protocol servers (/mcp, /mcp-admin) exposing 40+ tools and 11 React widgets to LLM clients via OAuth 2.1 + PKCE. The largest feature surface in the codebase.
• AI agent — a Pydantic AI conversational agent that streams SSE responses, calls into the same shared tool layer as the MCP servers, and supports both authenticated and guest modes.
• Blog — AI-generated, SEO-optimised blog content with a daily auto-publish scheduler. Perplexity powers generation; SEO fields are auto-computed from the post body.
• Notifications — multi-channel dispatch (push, email, SMS, in-app) driven by a central type registry, with per-user frequency caps and a shared APScheduler job.
• Vastu — a public, AI-powered floor plan analyzer that checks compliance with Vastu Shastra principles using the Gemini and GLM vision providers.

Navigation

Most feature pages follow the same structure: purpose, directory layout, key abstractions, a Mermaid diagram of the control flow, integration points, and entry points for modification. The MCP servers and Property Management pages are intentionally longer because of their surface area. For the building blocks underneath these features, see the systems pages, and for the core domain objects see the primitives pages.