All case studiesAI · Collaboration

Multi-Model AI Collaboration Platform

Workspace for comparing and orchestrating multiple LLM providers with real-time co-editing and RBAC.

  • Next.js
  • NestJS
  • WebSocket
  • Go
  • JWT
  • PostgreSQL

Screenshots

Neural FlowLive

Models

12

Jobs

248

Alerts

3

Problem

Teams wanted one place to run prompts across models—not separate tabs and undocumented prompt versions.

Latency-sensitive model routing and live co-editing had to coexist without overloading a single Node process.

Solution

NestJS owned auth, billing hooks, and provider adapters; Go handled throughput-sensitive orchestration.

WebSocket rooms synced document and chat state with explicit role checks on every action.

My role

Full-stack features—UI, APIs, real-time sessions, and integration contracts.

Architecture

  • Next.js front end on NestJS APIs
  • Go microservices for model orchestration
  • PostgreSQL for users, roles, and prompt metadata

Tech stack

  • Next.js
  • NestJS
  • WebSocket
  • Go
  • JWT
  • PostgreSQL

Client: Confidential · product startup

Impact

  • Single workspace for multi-provider experimentation
  • Governed access and rate limits per workspace
  • Versioned prompts with export for compliance review