Relationship intelligence productRebuilding

Orbit

A relationship operating system that turns professional contacts into a queryable, visual, and bilingual workspace.

RoleSolo product engineer

Timeline2025–present

Verified2026-08-11

A conceptual network view showing people, roles, and relationship strength without exposing real contacts.

Orbit / overview

The problem

Professional networks are usually scattered across contacts, notes, and memory, which makes follow-up inconsistent and relationship context difficult to query.

Database isolationtenant boundaries enforced with PostgreSQL RLSowner-verified evidence

Bilingual UXEnglish and Arabic interface with RTL supportowner-verified evidence

Rebuildingpublic landing page currently signals an offline productpublic evidence

Orbit / constraints

The constraints shaped the system

Users need fast relationship context without learning a CRM workflow.
Multi-tenant data isolation must hold at the database layer.
AI assistance must remain useful without becoming an opaque source of truth.

Orbit / decisions

The engineering decisions

RLS-first tenancy

Data isolation is enforced in PostgreSQL so the security boundary does not depend on every UI path remembering a filter.

Application-only tenant checks create too many places for a future feature to get wrong.

Graph plus semantic search

A force-directed relationship view gives visual structure while Gemini-backed search handles natural-language questions.

The graph is for orientation; search is for retrieval. They solve different jobs and should not be collapsed.

Billing as a system boundary

Stripe webhooks and server-side usage metering keep subscription state and limits out of client-only logic.

Payment events are external facts, so the application treats them as idempotent server-side inputs.

Orbit / Architecture

Architecture

Orbit combines a Next.js product surface with Supabase authentication and PostgreSQL RLS, a relationship graph, semantic search, Stripe billing, and bilingual interface primitives.

RLS protects tenant boundaries
Server-side billing and usage metering
Client cache hydration for responsive graph interactions

Orbit / Product and AI boundary

Product and AI boundary

The product uses AI to retrieve and enrich context, while the relationship model remains the durable source of truth. That separation keeps generated suggestions useful without making them authoritative records.

Orbit / What I would improve next

What I would improve next

Before relaunching publicly, I would complete a privacy review of imported contacts, add clearer consent and deletion affordances, and publish a verified production status.

Next case study

NexFlow

A lead pipeline platform that replaced conflicting spreadsheets across four business entities with one auditable, role-aware workflow.