Project Inventory

Complete inventory of projects in the ~/dev ecosystem, audited January 2026.

Summary Statistics

Core Platform

aicr (Monorepo)

• Location: ~/dev/aicr
• Purpose: Main platform monorepo
• Vercel: Yes (app.aicoderally.com)
• CLAUDE.md: Yes
• Structure:

  aicr/
  ├── apps/
  │   ├── aicr/        # Main app (:3000)
  │   ├── docs/        # Documentation (:3001)
  │   └── playground/  # API testing (:3002)
  ├── packages/
  │   ├── ai-router/   # AI routing
  │   ├── contracts/   # Shared types
  │   └── orbs/        # AI UI components
  └── services/
      ├── agent-conductor/  # AI orchestration
      └── sda-api/          # Document analysis

intelligentspm

• Location: ~/dev/intelligentspm
• Purpose: SPM knowledge platform with AskSPM RAG
• Vercel: Yes (intelligentspm.com (opens in a new tab))
• CLAUDE.md: Yes
• AI Integration:
  • Embedding: Ollama nomic-embed-text (768 dims)
  • LLM: Ollama llama3 / OpenAI gpt-4o-mini
  • Vector DB: pgvector
• Key Features:
  • 929 SPM knowledge cards
  • RAG-based Q&A (AskSPM)
  • Document analysis (via sda-core)

SPARCC Applications

The SPARCC suite consists of three separate applications that work together:

sgm-sparcc-demo (Sales Governance Manager)

• Location: ~/dev/demos/sgm-sparcc-demo
• Port: 3030
• Purpose: Sales compensation governance - policies, approvals, compliance
• Vercel: Yes
• CLAUDE.md: Yes
• AI Integration: AICR Gateway → Rally LLaMA → Claude API

sda-sparcc-demo (Sales Document Analyzer)

• Location: ~/dev/demos/sda-sparcc-demo
• Port: 3041
• Purpose: Document analysis against Gold Standard Library
• Vercel: Yes
• CLAUDE.md: Yes
• AI Integration: Multi-model consensus (via AICR Gateway)

scm-sparcc-demo (Sales Comp Manager)

• Location: ~/dev/demos/scm-sparcc-demo
• Port: 3042
• Purpose: Compensation plan CRUD, assignments, calculations
• Vercel: Yes (scm-sparcc-demo.vercel.app (opens in a new tab))
• CLAUDE.md: Yes
• AI Integration: AICR Gateway → Rally LLaMA → Claude API

Gold Standard Library

• Location: ~/dev/gold-standard-library
• Purpose: Canonical policy definitions
• Vercel: No (data only)
• CLAUDE.md: No
• Structure:

  gold-standard-library/
  ├── 02_POLICIES/           # 17 SCP policy JSONs
  │   └── _analysis_data/    # Requirement matrices
  ├── 03_TEMPLATES/          # DOCX templates
  └── 04_SAMPLES/            # Sample documents

AI Infrastructure

ai-router (Package)

• Location: ~/dev/aicr/packages/ai-router
• Purpose: Central AI routing
• Features:
  • Provider abstraction (OpenAI, Anthropic, Ollama)
  • Budget management
  • Response caching
  • Rate limiting

agent-conductor (Service)

• Location: ~/dev/aicr/services/agent-conductor
• Purpose: AI agent orchestration
• Version: 1.1 (with Ollama support)
• Documentation: Extensive (in service directory)

orbs (Package)

• Location: ~/dev/aicr/packages/orbs
• Purpose: Reusable AI UI components
• Components:
  • AskOrb - Chat interface
  • KBOrb - Knowledge base search

Client Projects (Sanitized)

BHG Demo

• Location: ~/dev/bhg-demo (if exists)
• Purpose: Client demonstration
• Vercel: Check project list

KPMG Demo

• Location: Referenced in docs
• Purpose: Enterprise demo
• Vercel: Check project list

Supporting Packages

sda-core

• Location: ~/dev/aicr/packages/sda-core
• Purpose: Document analysis engine
• Tests: 74 passing
• Features:
  • PDF/DOCX parsing
  • Gap analysis
  • Coverage scoring

contracts

• Location: ~/dev/aicr/packages/contracts
• Purpose: Shared TypeScript types
• Exports:
  • Context envelopes
  • Event schemas
  • Decision tokens

Project Status Matrix

Known Gaps

Environment Variables

• API keys fragmented across projects
• No central secrets management
• Manual propagation required

Documentation

• Several projects missing CLAUDE.md
• API references incomplete
• Runbooks outdated

Testing

• Some projects have no tests
• E2E coverage limited
• No integration test suite across projects

AI Integration

• Not all projects use ai-router
• Some have direct OpenAI calls
• Embedding dimensions inconsistent (768 vs 1536)

Recommendations

1. Centralize API Keys: Store in Vercel, propagate via env
2. Standardize AI Access: All projects use ai-router
3. Add CLAUDE.md: Every project needs one
4. Unify Embeddings: Pick 768 dims (Ollama-compatible)
5. Integration Tests: Add cross-project test suite