Text-to-JSON Structured Extractor

A versatile Apify actor that converts unstructured text and HTML into clean, structured JSON. Supports four extraction modes with auto-detection, URL fetching, and batch processing.

---

🎯 What It Does

Mode	Input	Output
Resume	Plain-text resume/CV	Contact info, experience, education, skills, certifications
E-Commerce	Product page HTML	Product name, price, brand, SKU, rating, images, availability
Blog SEO	Blog/article HTML	SEO score (A–F), meta tags, headings, links, content stats, recommendations
Chat Log	Chat exports (WhatsApp, Slack, Discord, IRC)	Messages, participants, topics, shared links, statistics
Auto	Any of the above	Detects the best extractor automatically

---

🚀 Quick Start

Minimal Input (uses defaults)

{
    "extractionMode": "auto",
    "inputType": "raw_text",
    "rawInput": "Jane Smith\njane@email.com\n\nExperience\nEngineer at Google..."
}

Fetch from URLs

{
    "extractionMode": "ecommerce",
    "inputType": "urls",
    "urls": [
        "https://example.com/products/widget-pro",
        "https://example.com/products/widget-lite"
    ],
    "outputFormat": "compact",
    "maxConcurrency": 10
}

Blog SEO Audit

{
    "extractionMode": "blog_seo",
    "inputType": "urls",
    "urls": ["https://myblog.com/latest-post"],
    "outputFormat": "full"
}

---

📥 Input Schema

Field	Type	Default	Description
extractionMode	enum	"auto"	resume, ecommerce, blog_seo, chat_log, or auto
inputType	enum	"raw_text"	raw_text, urls, or key_value_store
rawInput	string	(sample resume)	Direct text/HTML input
urls	string[]	[]	URLs to fetch content from
kvStoreKeys	string[]	[]	Keys to read from KV store
chatLogFormat	enum	"auto"	auto, whatsapp, slack, discord, irc, generic, simple
outputFormat	enum	"full"	full, compact, or flat
includeSourceText	boolean	false	Include original text in output
maxConcurrency	integer	5	Parallel URL fetches (1–20)
proxyConfiguration	object	Apify Proxy	Proxy settings for URL fetching
requestTimeoutSecs	integer	30	URL fetch timeout (5–120)

---

📤 Output Format

Each dataset record looks like:

{
    "source": "raw_input",
    "extraction_mode": "resume",
    "output_format": "full",
    "success": true,
    "error": null,
    "data": { ... }
}

Output Modes

• full — All extracted fields, deeply nested
• compact — Key fields only (great for dashboards)
• flat — Single-level dict with underscore-separated keys (great for spreadsheets)

---

🔍 Extraction Details

Resume Extractor

Detects and parses: name, email, phone, LinkedIn, GitHub, location, summary, work history with bullet points, education with GPA, categorized skills, projects, certifications, and languages.

E-Commerce Extractor

Three-priority pipeline: (1) JSON-LD Schema.org, (2) Open Graph meta tags, (3) HTML class-based parsing. Extracts product name, description, price with currency, brand, SKU, availability, rating, review count, and images.

Blog SEO Extractor

Produces a complete SEO audit with a score (0–100, grade A–F) based on 14 weighted checks. Analyzes title, meta description, Open Graph, Twitter Card, heading hierarchy, image alt text, internal/external links, structured data, content length, and more.

Chat Log Extractor

Auto-detects format from WhatsApp, Slack, Discord, IRC, and generic patterns. Builds participant profiles (message count, word average), extracts shared links with context, identifies topics via keyword frequency, and counts media messages.

---

🧪 Running Locally

# Install dependencies
pip install -r requirements.txt

# Run with Apify CLI
apify run --input-file=INPUT.json

---

📋 Example Output (Compact Resume)

{
    "name": "John Doe",
    "email": "johndoe@email.com",
    "phone": "(555) 123-4567",
    "location": "New York, NY",
    "summary": "Full-stack developer with 5+ years...",
    "skills": ["Python", "JavaScript", "TypeScript", "React", "Django", "AWS"],
    "experience_count": 2,
    "education_count": 1,
    "certifications": ["AWS Solutions Architect Associate", "Certified Kubernetes Administrator"],
    "languages": ["English (Native)", "French (Conversational)"]
}