Back to Projects
AI DevelopmentLiveWeb
NexumFive: Enterprise Multi-Model AI Chat Platform
A unified, credit-based AI workspace spanning a Web App with Split Mode and GrouPT, a Projects collaborative workspace with documents/spreadsheets/artifacts, an Admin Dashboard for marketplace and cost moderation, and a dual RAG layer (local FAISS plus a remote Milvus-backed microservice) — all driving five AI providers under one roof.
22+
Technologies
11+
Key Features
Web
Platform
Key Features
Multi-model AI chat with 5 providers (OpenAI, Claude, Gemini, Grok, Perplexity)Split Mode for side-by-side model comparisonGrouPT for multi-AI group discussionsDual RAG — local FAISS + remote Milvus with hybrid searchProjects collaborative workspace with TipTap and Fortune SheetArtifacts with HTML/JSX live preview and marketplaceGPTs/Experts marketplace with custom GPT builderWorkflows (Solutions) with autopilot and guest executionCredit-based billing via LemonSqueezy5-tier RBAC with AWS CognitoAdmin dashboard with per-provider cost analytics
Development Highlights
Unified 5-provider chat with streaming SSESplit Mode and GrouPT multi-AI discussionsDual RAG architecture — FAISS + Milvus with hybrid searchProjects workspace with TipTap and Fortune SheetWorkflows with autopilot and guest executionGPTs/Experts marketplace with custom GPT builder5-tier RBAC with per-provider cost analytics
Technology Stack
React 18TypeScriptViteTailwind CSSNode.jsExpressPrisma ORMPostgreSQLRedisBullMQOpenAIAnthropic ClaudeGoogle GeminixAI GrokPerplexityFAISSMilvus 2.6AWS CognitoAWS S3LemonSqueezyTipTapFortune Sheet
Our Contribution
Platform Architecture
Designed the five-surface platform architecture — Web App, Projects workspace, Admin Dashboard, RAG Microservice, and Local FAISS Service.
Unified Chat Layer
Built the unified multi-provider chat layer with streaming SSE responses across OpenAI, Claude, Gemini, Grok, and Perplexity from a single conversation thread.
Abilities System
Delivered the Abilities system — Perplexity-powered web search with citations, deep research sessions, RAG-context-injected image generation, file upload + RAG indexing, and the Nexum Five solution agent.
Split Mode & GrouPT
Implemented Split Mode and GrouPT — side-by-side dual-pane comparison and structured multi-AI discussions with debate, brainstorm, and analysis modes.
Artifacts System
Shipped the Artifacts system with HTML/JSX/TSX live preview, version history, conversation linking, social features, and a community marketplace.
Workflow Engine
Engineered the workflow engine (Solutions) with multi-step autopilot execution, progress bars, shareable links, and guest execution without authentication.
Projects Workspace
Built the Projects collaborative workspace combining TipTap document editor, Fortune Sheet spreadsheet editor, and an artifact code editor — all RAG-grounded against project files.
Dual RAG Architecture
Designed the dual-storage RAG architecture — local FAISS for per-conversation and per-GPT knowledge with disk-backed indices, and a remote Milvus-backed microservice for project-scale workloads.
Hybrid Retrieval
Implemented hybrid retrieval combining vector similarity, BM25 keyword matching, and recency scoring with IndexTag-scoped search for personalization.
Admin & Billing
Shipped the Admin Dashboard with user management, marketplace moderation, per-provider/model/user cost analytics, manual credit adjustments, pricing config, and full audit log. Integrated AWS Cognito authentication with 5-tier RBAC and the credit-based billing system on LemonSqueezy.
Results
AI Providers: Single interface for 5 leading AI modelsRAG Architecture: Production-grade dual RAG (FAISS + Milvus)Collaboration: RAG-grounded documents and spreadsheetsMarketplaces: Active GPTs, Workflows, and Artifacts marketplaces
What Makes This Special
Unifies five leading AI providers behind a single credit-based interface, adds collaborative project workspaces with grounded RAG, and introduces multi-AI features like Split Mode comparison and GrouPT group discussions that no single provider offers natively.
Frequently Asked Questions
MicrocosmWorks implemented a normalized credit system in NexumFive where each AI provider's token costs are mapped to a universal credit value maintained in the admin pricing configuration. When a user switches between OpenAI, Claude, Gemini, Grok, or Perplexity mid-conversation, the backend calculates credit consumption based on the per-model cost multiplier. The admin dashboard provides per-provider and per-model cost analytics for accurate margin tracking.
MicrocosmWorks designed the dual RAG system so local FAISS handles per-conversation and per-GPT knowledge with disk-backed indices for fast single-user retrieval, while the remote Milvus-backed microservice handles project-scale workloads requiring cross-document search across large file collections. Hybrid retrieval combines vector similarity, BM25 keyword matching, and recency scoring, with IndexTag-scoped search ensuring personalization within each user's data.
GrouPT is MicrocosmWorks' implementation of structured multi-AI discussions where multiple AI models participate in a conversation with defined roles. Users configure debate, brainstorm, or analysis modes, and the system orchestrates sequential responses from selected models (OpenAI, Claude, Gemini, Grok, Perplexity) via streaming SSE. Each model receives the full conversation context plus its role instructions, enabling comparative analysis of model strengths on the same prompt.
MicrocosmWorks built the Projects workspace combining TipTap for document editing, Fortune Sheet for spreadsheets, and an artifact code editor. All project files are automatically indexed into the Milvus RAG layer via a BullMQ embedding pipeline that processes documents on save. When users query the AI within a project context, retrieval is scoped to that project's indexed files, grounding responses in the team's actual documents and data.
MicrocosmWorks shipped a full marketplace system in NexumFive with community-published GPTs (custom AI personas with specialized instructions and knowledge), Workflows (multi-step automated processes with autopilot execution), and Artifacts (reusable HTML/JSX components with live preview). Users build custom GPTs through a no-code builder that configures system prompts, attached knowledge files indexed via FAISS, available tools, and model preferences.
Have a Similar Project in Mind?
Let's discuss how we can help you achieve similar results.