ContentMultiplier MVP

AI-powered content repurposing platform that transforms long-form content (webinars, podcasts, PDFs) into multi-channel marketing campaigns.

Key Features

• 4-Step Quality Pipeline: Atomize → Draft → Edit → Fact-check
• Content Library: Reusable atoms (quotes, stats, insights) that grow with each upload
• Persona-Driven: Tailor content to target audience priorities
• Universal Inputs: Video, audio, PDF, text
• Quality Scoring: Rank outputs by Hook, Brand Alignment, Clarity, Engagement
• Brand Voice Preview: Test persona transformations before processing
• Magic Words: Domain vocabulary for accurate transcription

Tech Stack

• Backend: Python FastAPI + SQLite
• AI Models:
  • Gemini 2.0 Flash (transcription, atomization, editing)
  • Claude Opus 4.5 (drafting with brand voice)
  • Gemini 2.5 Flash (fact-checking, scoring)
• Frontend: Vanilla JavaScript + Tailwind CSS
• Deployment: Replit (MVP), Railway/Render (production)

---

Quick Start

Prerequisites

• Python 3.13+
• API keys:
  • Anthropic (Claude) - console.anthropic.com
  • Google AI Studio (Gemini) - aistudio.google.com

Installation

1. Clone repository

   git clone [repo-url]
   cd contentmultiplier

2. Install dependencies

   pip install -r requirements.txt

3. Create environment file

   cp .env.example .env

4. Add your API keys to .env:

   ANTHROPIC_API_KEY=your_claude_key_here
   GEMINI_API_KEY=your_gemini_key_here
   SECRET_KEY=your_secret_key_here

5. Run the application

   python -m uvicorn app.main:app --reload

6. Access the app

   • Main app: http://localhost:8000
   • API docs: http://localhost:8000/docs
   • Admin panel: http://localhost:8000/admin/dashboard

---

First-Time Setup

1. Register an Account

• Go to http://localhost:8000/register.html
• Create account with email and password

2. Default Personas

Personas are pre-configured:

• CEO of Long-Term Care Facility - ROI-focused, executive level
• Director of Nursing (Memory Care) - Clinical, care-quality focused
• Marketing Director (Senior Living) - Brand and positioning focused

3. Test with Sample Content

1. Go to /upload.html
2. Select or paste sample content
3. Choose target persona
4. Enter magic words (optional): "YourBrand, Q4, ROI"
5. Select assets: 3 LinkedIn posts, 1 blog
6. Click "Upload & Process"
7. Watch status page for 10-15 minutes
8. View results and quality scores

---

Project Structure

contentmultiplier/
├── app/
│   ├── main.py                 # FastAPI application
│   ├── config.py               # Configuration settings
│   ├── database.py             # SQLite database setup
│   ├── api/                    # API endpoints
│   │   ├── auth.py             # Authentication (register, login)
│   │   ├── upload.py           # File upload
│   │   ├── jobs.py             # Job management
│   │   ├── library.py          # Content library
│   │   ├── personas.py         # Persona & brand voice preview
│   │   └── admin.py            # Admin endpoints
│   ├── services/               # Business logic
│   │   ├── transcription.py    # Gemini transcription
│   │   ├── atomization.py      # Content atomization
│   │   ├── drafting.py         # Content drafting (Claude)
│   │   ├── editing.py          # Audience editing
│   │   ├── factcheck.py        # Fact-checking
│   │   ├── scoring.py          # Quality scoring
│   │   └── pipeline.py         # Orchestrator
│   ├── models/                 # Pydantic models
│   ├── utils/                  # Utilities
│   │   ├── retry.py            # Retry logic
│   │   └── error_messages.py   # Enhanced error handling
│   └── templates/admin/        # Admin UI templates
├── data/
│   ├── prompts/                # Editable prompt templates
│   ├── clients/                # Client brand contexts
│   └── database/               # SQLite database
├── frontend/                   # User-facing UI
│   ├── index.html              # Landing page
│   ├── login.html              # Login
│   ├── register.html           # Registration
│   ├── upload.html             # Upload interface
│   ├── status.html             # Job status
│   ├── results.html            # Results display
│   ├── library.html            # Content library
│   ├── settings.html           # Settings & brand voice testing
│   └── static/                 # CSS, JS assets
├── tests/                      # Test suite
│   ├── test_e2e.py             # End-to-end tests
│   └── MANUAL_TEST_CHECKLIST.md
├── docs/                       # Documentation
│   └── BETA_ONBOARDING.md      # Beta user guide
└── scripts/                    # Utility scripts

---

Usage Guide

For Users

Upload Content:

1. Log in at /login.html
2. Go to Upload page
3. Upload video/audio/PDF or paste text
4. Select target persona
5. Choose output types (LinkedIn, Blog)
6. Add "magic words" for transcription accuracy (optional)
7. Click "Upload & Process"

View Results:

• See quality scores (0-100) for each output
• Hover over scores for breakdown (Hook, Brand, Clarity, Engagement)
• Copy or download content
• View extracted atoms

Use Content Library:

• Browse saved atoms from all uploads
• Filter by type (data, insight, story, problem, solution)
• Select atoms and generate new content without uploading

Test Brand Voice:

• Go to Settings page
• Enter sample text
• Select persona
• See side-by-side transformation preview

For Admins

Edit Prompts:

1. Go to /admin/prompts
2. Click "Edit" on any template
3. Modify prompt text
4. Save - changes apply to all new jobs

Monitor Usage:

• View job counts at /admin/dashboard
• Track API costs
• Review recent jobs

---

API Endpoints

Authentication

Method	Endpoint	Description
POST	/api/auth/register	Register new user
POST	/api/auth/login	Login
POST	/api/auth/logout	Logout

Jobs

Method	Endpoint	Description
POST	/api/upload	Upload content
GET	/api/job/{job_id}/status	Get job status
GET	/api/job/{job_id}/results	Get results

Library

Method	Endpoint	Description
GET	/api/library	Browse library
POST	/api/generate-from-library	Generate from atoms

Personas

Method	Endpoint	Description
GET	/api/personas	List personas
POST	/api/personas/preview-voice	Preview brand voice

Admin

Method	Endpoint	Description
GET	/admin/dashboard	Admin dashboard
GET	/admin/prompts	List prompts
PUT	/admin/prompts/{name}	Update prompt

---

Configuration

Environment Variables

Variable	Description	Required
GEMINI_API_KEY	Google Gemini API key	Yes
ANTHROPIC_API_KEY	Anthropic Claude API key	Yes
SECRET_KEY	JWT signing key	Yes
DATABASE_URL	SQLite database path	No (default: ./database/)
UPLOAD_MAX_SIZE_MB	Max upload size	No (default: 500)
ENVIRONMENT	development/production	No (default: development)

Prompt Templates

Edit in /data/prompts/ or via admin UI:

• atomization.txt - Content extraction
• linkedin_draft.txt - LinkedIn post generation
• blog_draft.txt - Blog article generation
• audience_edit.txt - Audience optimization
• factcheck.txt - Fact-checking

---

Cost Structure

Per job (typical 1-hour video):

Step	Model	Cost
Transcription	Gemini 2.0 Flash	~$0.05
Atomization	Gemini 2.0 Flash	~$0.10
Drafting	Claude Opus 4.5	~$2.00
Editing	Gemini 2.0 Flash	~$0.05
Fact-checking	Gemini 2.5 Flash	~$0.01
Scoring	Gemini 2.0 Flash	~$0.02
Total		~$2.23

---

Testing

Automated Tests

# Install test dependencies
pip install pytest pytest-asyncio httpx

# Run all tests
pytest tests/test_e2e.py -v --asyncio-mode=auto

# Run specific test class
pytest tests/test_e2e.py::TestUserJourney -v

Manual Testing

Follow the checklist in tests/MANUAL_TEST_CHECKLIST.md

---

Deployment

Replit (MVP)

1. Import repository to Replit
2. Set environment secrets in Replit
3. Click "Run"

See .replit and replit.nix for configuration.

Production (Railway/Render)

1. Connect GitHub repository
2. Set environment variables
3. Deploy

---

Troubleshooting

Jobs fail during transcription:

• Check API keys in .env
• Verify file is <500MB
• Check file format is supported (MP4, MP3, PDF, TXT)

Outputs seem generic:

• Review persona selection
• Edit prompt templates in admin
• Add magic words for domain vocabulary

Rate limit errors:

• Default: 10 uploads/hour/user
• Wait or adjust in app/main.py

Login issues:

• Clear browser storage
• Check token expiration (24 hours)

---

Development

Running Locally

python -m uvicorn app.main:app --reload --host 0.0.0.0 --port 8000

Adding New Personas

Edit app/services/persona_manager.py → get_default_personas()

Adding New Asset Types

1. Create prompt template in /data/prompts/
2. Add drafting function in app/services/drafting.py
3. Update pipeline in app/services/pipeline.py
4. Add UI elements in frontend

---

License

Proprietary - All rights reserved

Support

For issues: treynor@contentmultiplier.com