This Python project analyzes user prompts using fuzzy keyword matching to detect relevant tokens, brand terms, and keyword frequencies. It uses RapidFuzz for approximate string matching, nltk for multilingual stopword filtering, and supports both Markdown console output and structured JSON file export.

🔍 Features

Fuzzy matching using RapidFuzz
Multilingual stopword removal (English and German, configurable)
Token classification into brand and generic types
Flexible output format: markdown (printed to console) or json (saved to file)
Fully local, no Apify SDK required

📥 Example: Adidas Prompt Analysis

Here's a realistic input scenario for analyzing prompts that mention Adidas, including spelling variations and discount-related terms:

`input.json`

{
  "prompts": [
    "adidas sneaker sale today",
    "white running shoes from addidas",
    "Addidas Rabattcode gesucht",
    "neue sportschuhe im angebot",
    "discount für adidas laufschuhe",
    "adidas schuhe im trend",
    "sale bei addddidas online store",
    "welche adidas modelle sind beliebt",
    "adidass running collection deals"
  ],
  "target_keywords": ["adidas", "sale", "rabattcode", "angebot", "discount", "schuhe", "sneaker"],
  "brand_keywords": ["adidas", "addidas", "adidass", "addddidas"],
  "threshold": 65,
  "output_format": "markdown"
}

Sample Markdown Output:

token	matched_keyword	score	frequency	type
adidas	adidas	100	4	brand
addidas	adidas	91	1	brand
adidass	adidas	91	1	brand
rabattcode	rabattcode	100	1	generic
angebot	angebot	100	1	generic
discount	discount	100	1	generic
schuhe	schuhe	100	2	generic
sneaker	sneaker	100	1	generic

threshold: Minimum match score (0–100)
output_format: "markdown" (console) or "json" (file)

⚙️ Setup

Install dependencies:

pip install -r requirements.txt
python -m nltk.downloader punkt stopwords

Run the analyzer:

$python -m src

By default, it loads input.json and saves to output.json (if format is "json").

You can also specify custom paths using environment variables:

$APIFY_INPUT_PATH=custom_input.json OUTPUT_PATH=results.json python -m src

🧪 Output (Markdown Example)

| token    | matched_keyword | score | frequency | type    |
|----------|-----------------|-------|-----------|---------|
| purelei  | purelei         | 100   | 3         | brand   |
| rabatt   | rabattcode      | 86    | 2         | generic |
| discount | discount        | 100   | 1         | generic |

📝 Notes

Custom stopwords can be extended. please let me know via info@antonioblago.com
Compatible with Python 3.10+ or 3.11+

📚 Resources

On this page

Prompt Keyword Analyzer

Share Actor:

Website Content Crawler

apify/website-content-crawler

Crawl websites and extract text content to feed AI models, LLM applications, vector databases, or RAG pipelines. The Actor supports rich formatting using Markdown, cleans the HTML, downloads files, and integrates well with 🦜🔗 LangChain, LlamaIndex, and the wider LLM ecosystem.

Apify

55K

3.6

AI Linkedin Job Search

gokdeniz_kaymak/AI-Linkedin-Job-Search

A smart job search agent that analyzes your CV, scrapes LinkedIn job posts, and uses AI to filter and return only the most relevant opportunities.

Gökdeniz Kaymak

Google Maps Extractor

compass/google-maps-extractor

Extract data from hundreds of places fast. Scrape Google Maps by keyword, category, location, URLs & other filters. Get addresses, contact info, opening hours, popular times, prices, menus & more. Export scraped data, run the scraper via API, schedule and monitor runs, or integrate with other tools.

Compass

46K

4.4

AI Product Matcher

equidem/ai-product-matcher

Match products across multiple e-commerce websites. Use this AI product matching Actor whenever you need to find matching pairs of products from different online shops for dynamic pricing, competitor analysis or market research.

Matěj Sochor

588

Github Repo Markdown Scraper

louisdeconinck/github-repo-markdown-scraper

Transform GitHub repositories into a single, comprehensive markdown document effortlessly. Our tool streamlines analysis and processing, offering configurable file size limits, pattern filtering, and batch processing. Perfect for LLM AI prompts, it handles large repositories with ease.

Louis Deconinck

5.0

Amazon Bestsellers Scraper

junglee/amazon-bestsellers

Scrape the Amazon Best Sellers categories and extract details on top 100 most popular items on Amazon. Download product name, price, URL, and thumbnail image. Best used on .com, .co.uk, .de, .fr, .es, and .it domains. Download your data in various formats: HTML table, JSON, CSV, Excel, and more.

Junglee

1.2K

4.3

AI LinkedIn Job Matcher

james.logantech/ai-linkedin-job-matcher

AI LinkedIn Job Matcher helps job seekers find the most relevant LinkedIn job postings using NLP, and OpenAI's GPT-4. It analyzes job descriptions, matches them to resumes, and ranks opportunities by relevance. Automate job searching, save time and discover the best career matches easily!

James

Craiyon AI Image Creator (DALL·E mini)

muhammetakkurtt/craiyon-ai-image-creator

Craiyon AI Image Creator is an Apify actor that generates AI images from prompts. Users can exclude unwanted elements, select generation types (Photo, Drawing, Vector), choose aspect ratios (Square, Portrait, Landscape). It offers fast and customizable image creation with concurrent request handling

Muhammet Akkurt

113

5.0

AI Content Topic Generator 🎯

easyapi/ai-content-topic-generator

🚀 Generate trending content ideas and topics based on keywords! Get AI-powered suggestions with SEO benefits analysis and relevance explanations. Perfect for content creators, marketers, and SEO specialists looking to boost engagement and search rankings. ✨

EasyApi

Google-LLM Image Search Scraper

loongnian714/google-llm-image-search-scraper

Just describe what you need—our AI interprets your prompt and searches from every angle to surface the most relevant, non-AI-generated visuals. Perfect for everyone who need trustworthy, real imagery fast. With multi-perspective analysis,find up to 4x more relevant results than basic image searches.