PersonaPost 🤖✨

Analiza el estilo de comunicación de LinkedIn y genera publicaciones auténticas con IA

PersonaPost scrapes a LinkedIn user's public posts, uses Google Gemini 1.5 Pro to build a detailed personality profile, and then generates 20 LinkedIn posts on any topic — written exactly as that person would write them.

---

Tech Stack

Layer	Technology
Frontend	React 18 + Vite + TailwindCSS + Zustand
Backend	Python 3.11 + Flask 3 + SQLAlchemy
Database	SQLite (embedded)
AI	Google Gemini 1.5 Pro API
Scraping	Playwright (headless Chromium)

---

Quick Start

Prerequisites

• Python 3.10+
• Node.js 18+
• Google Gemini API Key — Get one here

---

1. Backend Setup

cd personapost/backend

# Create virtual environment
python -m venv venv
venv\Scripts\activate       # Windows
# source venv/bin/activate  # macOS/Linux

# Install dependencies
pip install -r requirements.txt

# Install Playwright browser
playwright install chromium

# Copy environment template
copy .env.example .env

Edit .env and fill in your values:

GEMINI_API_KEY=your_actual_gemini_api_key
FLASK_SECRET_KEY=any_random_string_here

Start the Flask server:

python app.py
# Server runs on http://localhost:5000

---

2. Frontend Setup

cd personapost/frontend

# Install dependencies
npm install

# Start dev server
npm run dev
# App runs on http://localhost:5173

---

Using the App

Flow 1 — Analyze a LinkedIn Profile

1. Go to Analizar page
2. Paste a LinkedIn URL (e.g. https://www.linkedin.com/in/midudev/)
3. Click Analizar Perfil — watch the progress steps
4. View the Personality Card with tone, technical level, emoji usage, key phrases

Scraping tip: LinkedIn requires authentication for many profiles. If scraping fails, use the "Pegar publicaciones manualmente" option and paste the raw post texts separated by ---.

Flow 2 — Generate Posts

1. Go to Generar page
2. Select an analyzed profile from the dropdown
3. Enter a topic (e.g. "IA en desarrollo web", "manga y anime")
4. Click Generar 20 Publicaciones
5. Wait 30–60 seconds while Gemini generates your posts
6. Copy individual posts or browse all in the grid

Flow 3 — History

• View all analyzed profiles in Historial
• Browse past batches and individual posts
• Delete profiles or specific posts

---

Environment Variables

Variable	Required	Description
GEMINI_API_KEY	✅	Your Google Gemini API key
FLASK_SECRET_KEY	✅	Random secret for Flask sessions
DATABASE_URL	❌	SQLite path (default: sqlite:///personapost.db)
LINKEDIN_EMAIL	❌	LinkedIn email for authenticated scraping
LINKEDIN_PASSWORD	❌	LinkedIn password for authenticated scraping

---

API Reference

Method	Endpoint	Description
POST	/api/analyze-profile	Scrape + analyze a LinkedIn profile
GET	/api/profiles	List all analyzed profiles
GET	/api/profiles/<id>	Get profile with full analysis
DELETE	/api/profiles/<id>	Delete profile + all data
POST	/api/generate-posts	Generate 20 posts for a topic
GET	/api/batches/<profile_id>	Get all batches for a profile
GET	/api/posts/<batch_id>	Get all posts in a batch
DELETE	/api/posts/<id>	Delete a single post
PATCH	/api/posts/<id>/copy	Increment copy counter

---

Project Structure

personapost/
├── backend/
│   ├── app.py              Flask application factory
│   ├── config.py           Configuration from .env
│   ├── models.py           SQLAlchemy ORM models
│   ├── routes/
│   │   ├── profiles.py     Profile analysis endpoints
│   │   └── posts.py        Post generation endpoints
│   ├── services/
│   │   ├── scraper.py      LinkedIn Playwright scraper
│   │   ├── gemini.py       Gemini API integration
│   │   └── analyzer.py     Analysis orchestrator
│   ├── requirements.txt
│   └── .env.example
├── frontend/
│   ├── src/
│   │   ├── components/
│   │   │   ├── Layout/     Navbar, Layout wrapper
│   │   │   ├── Analyzer/   ScrapeProgress, PersonalityCard
│   │   │   ├── Generator/  PostCard, PostGrid
│   │   │   └── History/    ProfileList
│   │   ├── pages/
│   │   │   ├── Home.jsx    Landing page
│   │   │   ├── Analyze.jsx LinkedIn URL + analysis flow
│   │   │   ├── Generate.jsx Post generation
│   │   │   ├── Posts.jsx   View generated posts
│   │   │   └── History.jsx Past profiles + batches
│   │   ├── hooks/
│   │   │   ├── useProfileAnalysis.js
│   │   │   └── usePostGenerator.js
│   │   ├── services/api.js All axios calls
│   │   └── store/appStore.js Zustand global state
│   ├── package.json
│   └── vite.config.js
└── README.md

---

Notes on LinkedIn Scraping

LinkedIn actively blocks scraping. Here's what PersonaPost does to mitigate:

• ✅ Randomized delays (2–5 seconds between actions)
• ✅ Realistic User-Agent rotation
• ✅ Automation flag masking (navigator.webdriver = undefined)
• ✅ Optional authenticated scraping with your credentials
• ✅ Manual paste fallback — if scraping fails, simply paste post texts

If scraping consistently fails, use the manual paste mode — it works identically for the AI analysis.