EXACT API ARCHITECTURE (OPENAI + SUPABASE)

1. Core Stack

• Frontend: React, Next.js, Tailwind, Zustand or React Context
• Backend: Next.js API routes or Supabase Edge Functions
• Database: Supabase Postgres
• Auth: Supabase Auth
• AI: OpenAI API
• Storage: Supabase Storage

2. Database Schema (Supabase)

• `users`: id, email, created_at, plan, full_name
• `claims`: id, user_id, claim_name, status, created_at, updated_at
• `claimant_profiles`: id, claim_id, conditions, symptoms, aids, assistance, risks, bad_day_summary, typical_day_summary, supporting_evidence
• `activities`: id, claim_id, activity_code, activity_name, raw_answer, rewritten_answer, descriptor_code, descriptor_text, points, confidence, evidence_gaps, follow_up_questions
• `mock_assessments`: id, claim_id, mode, status, total_score, summary
• `mock_assessment_messages`: id, mock_assessment_id, role, message, scoring
• `generated_documents`: id, claim_id, document_type, title, content, file_url
• `evidence_uploads`: id, claim_id, file_name, file_url, file_type

3. Core API Endpoints

1. `POST /api/claims`: Create claim 2. `POST /api/claims/:claimId/intake`: Save intake 3. `POST /api/claims/:claimId/follow-up`: Generate follow-up questions 4. `POST /api/claims/:claimId/rewrite-activity`: Rewrite activity answer 5. `POST /api/claims/:claimId/score`: Score full claim 6. `POST /api/claims/:claimId/consistency-check`: Run consistency check 7. `POST /api/claims/:claimId/mock-assessment/start`: Start mock assessment 8. `POST /api/mock-assessment/:id/message`: Send mock assessment answer 9. `POST /api/mock-assessment/:id/report`: Generate coaching report 10. `POST /api/claims/:claimId/generate-mr`: Generate MR letter

4. MVP Build Order

Phase 1

• Auth
• Claim creation
• Intake form
• Activity answer rewriting
• Save/load claim state

Phase 2

• Descriptor scoring
• Estimated totals
• Upgrade/paywall flow

Phase 3

• Mock assessment
• Coaching feedback
• Consistency checker

Phase 4

• Evidence builder
• MR document generator
• Exportable claim pack