Propy — Client Intelligence Dashboard

An internal agency tool that scrapes client websites, detects their tech stack, researches competitors, and uses AI to generate personalized automation proposals — viewable in-dashboard and exportable as PDF.

---

What it does

For each client, Propy runs a 7-step analysis pipeline:

1. Scrape — fetches the client's website and extracts content
2. Surface tech detection — identifies CMS, analytics, ad pixels, SaaS tools
3. Deep tech detection (on demand) — fingerprints frameworks, CRM, payments, live chat
4. Company research — web search summary of the company and industry
5. Competitor discovery — auto-discovers competitors + merges manual overrides
6. AI proposal — generates a structured proposal with pain points, tool recommendations, and ROI estimates
7. PDF export — renders the proposal as a styled, downloadable PDF

The AI layer is model-agnostic — switch between Claude and OpenAI via a single environment variable.

---

Tech Stack

Layer	Technology
Frontend	React 18 + TypeScript + Vite + Tailwind CSS
Backend	FastAPI (Python 3.12)
Task Queue	Celery 5 + Redis 7
Database	PostgreSQL 16
ORM	SQLAlchemy 2 (async) + Alembic
PDF Export	Playwright (Chromium)
Web Search	Tavily
LLM	Claude (default) or OpenAI GPT-4o
Containers	Docker + Docker Compose
Reverse Proxy	Nginx (production)

---

Project Structure

propy/
├── backend/
│   ├── app/
│   │   ├── main.py              # FastAPI app + routers
│   │   ├── config.py            # Environment settings
│   │   ├── database.py          # Async SQLAlchemy setup
│   │   ├── llm/
│   │   │   └── adapter.py       # Model-agnostic LLM adapter
│   │   ├── models/              # SQLAlchemy ORM models
│   │   ├── schemas/             # Pydantic request/response schemas
│   │   ├── routers/             # API route handlers
│   │   └── workers/
│   │       ├── celery_app.py    # Celery configuration
│   │       ├── tasks.py         # Async task definitions
│   │       └── pipeline.py      # 7-step pipeline logic
│   ├── alembic/                 # Database migrations
│   └── requirements.txt
├── frontend/
│   ├── src/
│   │   ├── pages/               # ClientsPage, ClientDetailPage, RunDetailPage
│   │   ├── components/          # Layout, StatusBadge
│   │   └── lib/                 # API client + types, utilities
│   └── nginx.conf               # Nginx config for SPA
├── nginx/
│   └── nginx.conf               # Reverse proxy + SSL (production)
├── scripts/
│   ├── deploy.sh                # One-command VPS deploy
│   └── backup.sh                # Daily pg_dump cron script
├── storage/reports/             # Generated PDFs
├── docker-compose.yml           # Development stack
├── docker-compose.prod.yml      # Production stack
└── .env.example                 # Environment variable template

---

Getting Started (Local)

Prerequisites

• Docker + Docker Compose
• API keys: Anthropic or OpenAI, and Tavily (for web search)

Setup

git clone git@github.com:Danultimate/propy.git
cd propy

cp .env.example .env
# Edit .env with your API keys

Start

docker compose up -d

Service	URL
Dashboard	http://localhost:3000
API	http://localhost:8000
API Docs	http://localhost:8000/docs

Run database migrations

docker compose exec backend alembic upgrade head

Logs

docker compose logs -f backend   # FastAPI
docker compose logs -f worker    # Celery pipeline

Stop

docker compose down
# Add -v to also remove volumes (full reset)

---

Environment Variables

# Database
POSTGRES_USER=propy
POSTGRES_PASSWORD=changeme
POSTGRES_DB=propy
DATABASE_URL=postgresql+asyncpg://propy:changeme@db:5432/propy

# Redis
REDIS_URL=redis://redis:6379/0

# LLM — choose one
LLM_PROVIDER=claude        # claude | openai
ANTHROPIC_API_KEY=
OPENAI_API_KEY=

# Web Search
TAVILY_API_KEY=

# App
SECRET_KEY=change-this-in-production
ENVIRONMENT=development    # development | production

---

API Reference

Clients

Method	Endpoint	Description
GET	/api/clients	List all clients
POST	/api/clients	Create client
GET	/api/clients/:id	Get client
PATCH	/api/clients/:id	Update client
DELETE	/api/clients/:id	Delete client

Analysis Runs

Method	Endpoint	Description
POST	/api/clients/:id/runs	Trigger analysis
GET	/api/clients/:id/runs	List runs
GET	/api/clients/:id/runs/:runId	Get run + result
POST	/api/clients/:id/runs/:runId/regenerate	Regenerate proposal only
GET	/api/clients/:id/runs/:runId/pdf	Download PDF

Competitors (manual overrides)

Method	Endpoint	Description
GET	/api/clients/:id/competitors	List manual competitors
POST	/api/clients/:id/competitors	Add competitor
DELETE	/api/clients/:id/competitors/:cId	Remove competitor

Config

Method	Endpoint	Description
GET	/api/config/providers	List LLM providers
PATCH	/api/config/provider	Switch active provider

---

Proposal Structure

Every AI-generated proposal follows this structure:

1. Executive Summary — estimated hours/month saved and annual cost reduction
2. Company & Tech Snapshot — what they do, tech stack mapped as gaps
3. Pain Points — 3–5 evidence-backed, competitor-informed findings
4. Recommended Tools — tool name, what it solves, effort, estimated cost
5. ROI Estimates — hours saved + industry benchmarks per recommendation, summary table
6. Next Steps — 3 clear call-to-action bullets

---

Deploying to a VPS

1. Prepare the server

# On your VPS
mkdir -p /docker/propy && cd /docker/propy
git clone git@github.com:Danultimate/propy.git .
cp .env.example .env
# Fill in production values — strong passwords, real API keys

2. SSL certificates

Place Let's Encrypt certificates at:

nginx/certs/live/your-domain.com/fullchain.pem
nginx/certs/live/your-domain.com/privkey.pem

Update nginx/nginx.conf — replace your-domain.com with your actual domain.

3. Deploy

bash scripts/deploy.sh

This pulls the latest code, builds images, runs migrations, and starts all services.

4. Set up daily backups

# Add to crontab (crontab -e)
0 2 * * * /docker/propy/scripts/backup.sh

Backups are stored at /docker/propy/backups/ and kept for 14 days.

Update

bash scripts/deploy.sh

---

Database Schema

clients              — client records
analysis_runs        — pipeline execution history (status, timestamps)
analysis_results     — pipeline output (tech_stack JSON, proposal, pdf_path)
competitors_override — manually added competitors per client

Run statuses: pending → running → completed | failed

---

Switching LLM Provider

Via the dashboard config panel, or directly:

# Switch to OpenAI
curl -X PATCH http://localhost:8000/api/config/provider \
  -H "Content-Type: application/json" \
  -d '{"provider": "openai"}'

Make sure the corresponding API key is set in .env.