Bibek Khatiwada SEO Strategist
WhatsApp
Tools Showcase

Open-Source & Custom SEO Systems Showcase.

Give tangible form to the Computer Science & Automation Engineer identity. Explore Python NLP scripts, n8n automated workflow pipelines, and custom static generators.

NLP & Semantic Pipeline

Content Semantic Flow Checker

spaCy + sentence-transformers (all-MiniLM-L6-v2) pipeline for analyzing semantic similarity, topic drift, and entity coverage between content headers and search intent.

Technologies
Python 3.11spaCysentence-transformersPyTorchStreamlit
Key Capabilities
  • Cosine similarity matrix across document headings
  • Entity extraction and Wikidata link mapping
  • Sub-topic coverage scoring against top-ranking SERP competitors
  • Automated suggestion engine for missing semantic attributes
GitHub Repo →
Search Automation

n8n & Colab GSC Anomaly Detector

Automated workflow engine monitoring daily Google Search Console API data for unexpected click/impression drops, cannibalization flags, and SERP layout shifts.

Technologies
n8nPythonGoogle ColabSearch Console APISlack Webhooks
Key Capabilities
  • Daily statistical Z-score anomaly alerts sent to Slack/Email
  • Page-level vs query-level drop isolation
  • Automated Google Sheets dashboard sync via Apps Script
  • AI Overview zero-click loss tracking
GitHub Repo →
AI Search Readiness

Automated llms.txt & Markdown Generator

Static generator module that automatically parses HTML site structure and builds a standardized, clean markdown /llms.txt file for AI web crawlers.

Technologies
PythonBeautifulSoup4Jinja2Markdown
Key Capabilities
  • Automatic extraction of core site entities and case study metrics
  • Strip out tracking scripts, CSS, and navigation bloat for clean LLM ingestion
  • Hierarchical markdown formatting with canonical link attribution
  • Seamless integration with Python static site build scripts
GitHub Repo →
Live Interactive Demo

Test Semantic Vector Scoring in Your Browser

Type a target intent query and three candidate headings. Our client-side algorithm simulates the spaCy + Sentence-Transformers cosine distance check to demonstrate which headings survive LLM context pruning.

Backed by spaCy + sentence-transformers algorithm
Generative Engine Optimization

Live llms.txt File Generator & Validator

Generate a clean, standardized /llms.txt file configured for AI crawlers like Anthropic, Perplexity, and OpenAI.

PREVIEW: /llms.txt
Click "Generate Validated llms.txt" above to generate your crawler file...
Production Telemetry Sandbox

GSC Statistical Anomaly & Algorithmic Drop Detector

Simulate how our BigQuery + GSC API pipeline detects algorithmic penalty anomalies using standard deviation Z-scores rather than noisy day-over-day vanity metrics.

Statistical Significance
Z = -3.57σ
Volume Deviation
-1,000 clicks (-31.2%)
CRITICAL ANOMALY DETECTED
SIMULATED PAGERDUTY / SLACK WEBHOOK PAYLOAD (JSON)

      

    

  







  

    

      

        Enterprise Crawler Governance
        
AI Crawler Robots Directives & Scraping Configurator

        
Configure granular crawler permissions for search-engine retrieval models vs non-consensual AI training scrapers. Generates a standard-compliant robots.txt snippet.

      

    


    

      

        
        
        
        
        
        
      


      

        

          
GENERATED ROBOTS.TXT SNIPPET

          
        

        

      

    

  








  

    

      

        Production Python Code
        
The Core NLP Pipeline Behind the Algorithm

        
A look at the production code from Bibek's public GitHub repository Content---Semantic-Flow-Checker-.

      

    

    

      

        
        semantic_flow_checker.py · GitHub @bibek1223
      

      
from sentence_transformers import SentenceTransformer, util
import spacy

# Load lightweight transformer model and spaCy parser
nlp = spacy.load("en_core_web_sm")
model = SentenceTransformer("all-MiniLM-L6-v2")

def analyze_semantic_flow(query: str, headers: list[str]):
    query_emb = model.encode(query, convert_to_tensor=True)
    header_embs = model.encode(headers, convert_to_tensor=True)
    
    # Calculate cosine similarity matrix
    scores = util.cos_sim(query_emb, header_embs)[0]
    
    return [
        {"header": h, "similarity_score": round(float(score), 4)}
        for h, score in zip(headers, scores)
    ]

# Example Output
# Header 1: 0.8421 (High Salience · RAG Retained)
# Header 2: 0.7914 (High Salience · RAG Retained)
# Header 3: 0.1102 (Off-topic Drift · Pruned by LLM Retriever)

    

  




    

      

        

          let's talk
          
Need a custom SEO automation tool or crawler?

          
Tell me your manual reporting or auditing bottleneck and I will build an automated workflow.

        

        

          Write me a brief