Nexrena Platform · CRM, Forms, and AI Ops System
Next.js, Express, PostgreSQL platform with multi-site forms/chat, Gemini sales assistant, OpenAI copilot, and client portal.
Lead engineer · sole builder and operator of the platform
1
Unified ops system
5+
SaaS tools replaced
CRM
PM · invoicing · APIs
Live
Daily production use
Nexrena is my client studio for B2B web design and SEO. Running it on scattered SaaS subscriptions · CRM here, invoicing there, project tracking in a third tool · created friction on every client handoff. I built a single platform to own the full delivery loop.
The problem
Agency operations were spread across:
- CRM and lead tracking in one product.
- Project management and task status in another.
- Invoicing, subscriptions, and payment follow-up in a third.
- Client communication and file handoffs in ad hoc channels.
Every new client meant context switching across tools, manual reconciliation, and no single source of truth for delivery status.
My role
I’m the only engineer. I designed it, built it, and run it every day.
What I built
Next.js client portal and admin UI
- Client-facing views for project status, deliverables, and service upgrades.
- Internal admin for leads, engagements, and delivery tracking.
- Authentication and role separation between client and operator workflows.
Express API layer
- REST endpoints for CRM records, project milestones, invoicing events, and subscription state.
- Validation and error handlingbuilt for real client data.
- Heroku-deployed backend with Postgres persistence.
PostgreSQL data model (Prisma)
- Unified schema for contacts, projects, invoices, subscriptions, and service catalog items.
- Relational integrity across entities that previously lived in separate SaaS silos.
- Migration workflow for schema changes without downtime on live client data.
Integrated client workflows
- Service upgrade paths (pages, SEO, forms, automations) connected to the same project record.
- Multi-site public form API with site-key routing, sanitization, dedupe, and ops inbox notifications.
- AI chat intake for client sites with site-specific knowledge profiles and transcript analytics.
Practical AI layer
- Sales assistant · Gemini-backed chat with lexical retrieval, guardrails, lead scoring, and Postgres session analytics.
- Ops/client copilot · OpenAI tool-calling over live CRM/inbox data with scoped writes and confirmation flow.
- Portfolio recruiting assistant · grounded chat with retrieval evals (270+ questions) on this site.
Outcomes
| Result | Detail |
|---|---|
| Tool consolidation | CRM, PM, invoicing, and client portal in one system |
| Daily use | Platform runs live client delivery today |
| Cost | Eliminated multiple SaaS seats for core ops |
| Extensibility | New service types added as schema + UI modules, not new vendors |
Same approach I use on client work · own the data model, ship the API, and keep it running in production.
What this demonstrates
- Full-stack ownership from schema design through deployed UI.
- Building internal platforms that replace SaaS sprawl · relevant for product engineering and platform teams.
- Shipping and operating software you depend on daily, not portfolio-only demos.
Live system
www.nexrena.com ↗