AI assistant for real estate leads
on WhatsApp
Alloha is a private MVP that explores AI-powered lead qualification, property search, and broker handoff using FastAPI, Next.js, Supabase/PostgreSQL, Redis, and WhatsApp-oriented workflows.
Evidence
Built as a technical case study.
The landing now reflects the current engineering state: useful MVP architecture, clear gaps, and no inflated product claims.
330 active listings
Measured in Supabase during the May 15, 2026 technical audit.
RAG-ready, not RAG-live
pgvector/RPC 384d and RAG code exist, but the primary endpoints still use lexical search.
Known contract gap
The API currently needs schema alignment before public production use.
Pilot flow
From inquiry to broker handoff.
A practical workflow for testing AI-assisted real estate lead handling.
Capture lead intent
Collect budget, location, property type, timing, and contact context from WhatsApp-style conversations.
Search property data
Query the property database with filters today, with pgvector/RAG prepared as the next retrieval layer.
Prepare handoff
Summarize lead qualification and candidate listings so a broker can follow up with context.
Property search
Filter listings by intent, neighborhood, price, and bedrooms.
Broker handoff
Prepare structured lead context for a human follow-up.
WhatsApp-oriented flow
Designed around conversational lead intake and response.
RAG-ready architecture
Supabase/PostgreSQL with pgvector schema and RPC are included for semantic retrieval work.
Dashboard path
Frontend routes exist for onboarding, setup, login, and operational screens.
What is implemented
Honest MVP scope.
Alloha is positioned as a private technical MVP for validating real estate AI workflows, not as a scaled SaaS with invented traction.
FastAPI backend
Canonical endpoints for chat, listing search, ingest, auth, onboarding, leads, and system status.
Supabase data model
PostgreSQL schema includes properties, conversations, messages, leads, idempotency, and pgvector-ready fields.
Model gateway
OpenRouter-oriented gateway with hard-stop behavior, rate limits, and Redis-backed metrics hooks.
Current audit status
The repo demonstrates a credible engineering direction, while the live data path still needs API/schema alignment and embedding backfill before semantic search claims are fair.
Private MVP - May 2026 audit
Technical features
Designed for real estate workflows.
A focused stack for lead qualification, property search, and future CRM-ready operations.
Lead qualification
Extract budget, urgency, bedrooms, location, timing, and contact readiness from conversations.
Property search API
Lexical and filter-based listing search is implemented as the current endpoint path.
Guardrailed model gateway
Rate limits and hard-stop fallback avoid pretending the model layer is always available.
pgvector schema
The repo includes vector(384) schema, indexes, and RPC for semantic retrieval work.
Ingest workflow
Official feed first, scraper fallback, locking, change detection, and deactivation logic.
Pilot dashboard
Next.js routes support login, onboarding, setup, dashboard, and contact flows.
Coverage
A practical MVP surface for pilots.
The product story is now limited to capabilities represented in the repository or audit findings.
Pilot access
Private MVP, not public SaaS.
Alloha is available as a technical case study and pilot conversation while core retrieval work is being completed.
For recruiters, technical reviewers, and selected real estate pilot conversations.
- Private demo of the Next.js and FastAPI workflow
- Architecture walkthrough: Supabase, pgvector, Redis, WhatsApp flow
- Clear disclosure of current gaps and roadmap
- GitHub-visible technical case study
- No invented customer logos or vanity metrics
- Roadmap for embeddings, RAG, analytics, CRM, voice, and multi-tenant support
- Pilot discussion after API/schema contract fixes
FAQ
Frequently asked questions.
Straight answers about what Alloha is today.
No. Alloha is a private MVP and technical case study. It is suitable for portfolio review and selected pilot conversations, not public production claims.
Review the case study.
See the architecture, audit findings, implemented workflow, and next technical steps without inflated traction claims.
Private MVP - RAG-ready architecture - public-production gaps disclosed