Your Primary Assignment: Expand AI is built on the WorkAlly engine with specialized automation for proposals and 3D generation.
Overview
Expand AI provides full-cycle automation for RFPs and 2D/3D proposal generation, leveraging the WorkAlly intelligence engine.
Proposal Generation
Automated content and 2D design creation
3D Module
Blender MCP orchestration for 3D assets
Opportunity Intel
Market intelligence and opportunity scoring
Financial Proposals
Automated pricing and submission flows
Tech Stack
Core
AI Components
Infrastructure
const expandStack = {
frontend: "Next.js 14 App Router",
styling: "Tailwind CSS + shadcn/ui",
backend: "Supabase (Postgres, Auth, Storage)",
ai: "LangChain + LangGraph",
deployment: "Vercel + Supabase Edge Functions"
}
- WorkAlly Engine: Core intelligence layer
- LangChain: Orchestration and tool calling
- LangGraph: Complex workflow management
- Together AI: Inference and fine-tuning
- Blender MCP: 3D generation module
- CI/CD: GitHub Actions
- Monitoring: Sentry
- Analytics: PostHog
- Storage: Supabase Storage + S3
- CDN: Cloudflare
Key Features
Proposal Generation
RFP Analysis
AI extracts requirements, scoring criteria, and submission guidelines
Content Generation
Creates tailored responses using company knowledge base
Design Assembly
Generates professional 2D layouts with branding
Review & Edit
Human-in-the-loop editing with AI suggestions
Submission
Automated formatting and submission workflows
3D Module Orchestration
- Procedural 3D model generation
- Texture and material application
- Scene composition
- Rendering pipelines
- Asset optimization
// Example Blender MCP integration
const generate3DAsset = async (prompt: string) => {
const response = await blenderMCP.generate({
prompt,
style: "photorealistic",
format: "glb",
optimization: "web"
})
return response.assetUrl
}
Development Guide
Getting Started
Clone Repository
git clone https://github.com/Weventures-AI/expand-ai.git
cd expand-ai
Environment Setup
cp .env.example .env.local
# Add your Supabase and AI keys
Read Driver Docs
Review claude.md, plan.md, and agents.md
Working with AI Workflows
// app/api/proposals/generate/route.ts
import { LangChain } from '@langchain/core'
import { proposalChain } from '@/lib/chains'
export async function POST(request: Request) {
const { rfpContent, companyContext } = await request.json()
const proposal = await proposalChain.invoke({
rfp: rfpContent,
context: companyContext,
tools: ['research', 'write', 'format']
})
return Response.json(proposal)
}
Database Schema
-- Proposals table
CREATE TABLE proposals (
id UUID PRIMARY KEY,
rfp_id UUID REFERENCES rfps(id),
content JSONB,
status TEXT,
created_at TIMESTAMPTZ
);
-- RFPs table
CREATE TABLE rfps (
id UUID PRIMARY KEY,
title TEXT,
requirements JSONB,
deadline TIMESTAMPTZ,
score_criteria JSONB
);
-- Assets table
CREATE TABLE assets (
id UUID PRIMARY KEY,
proposal_id UUID REFERENCES proposals(id),
type TEXT, -- '2d', '3d', 'document'
url TEXT,
metadata JSONB
);
Current Milestones
Q1 2024
Q2 2024
Current Sprint
- Implementing LangGraph workflows
- Supabase RLS policies
- Edge function optimization
- Performance improvements
Testing
pnpm test # Run all tests
pnpm test:watch # Watch mode
pnpm test:coverage # Coverage report
Deployment
Environments
Development
Local development with hot reload
Staging
Preview deployments on Vercel
Production
Production on custom domain
CI/CD Pipeline
Support & Resources
Next Steps: Start with the setup guide and review the driver documents in the repository. 
