Deploying an NLP Chatbot on Render with Flask

What I Built

A simple but effective FAQ chatbot that understands natural language questions and returns the most relevant answer from a knowledge base. No APIs, no cloud AI services — just pure Python NLP.

The NLP Pipeline

1. Tokenization — Break user input into words using NLTK
2. Vectorization — Convert tokens to TF-IDF vectors
3. Similarity — Compute cosine similarity against all FAQ entries
4. Response — Return the best match above a threshold

from nltk.tokenize import word_tokenize
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.metrics.pairwise import cosine_similarity

def get_response(user_input, faq_data):
    tokens = word_tokenize(user_input.lower())
    processed = " ".join(tokens)
    
    all_texts = [processed] + [item["question"] for item in faq_data]
    tfidf = TfidfVectorizer()
    vectors = tfidf.fit_transform(all_texts)
    
    similarities = cosine_similarity(vectors[0:1], vectors[1:])
    best_idx = similarities.argmax()
    
    if similarities[0][best_idx] > 0.3:
        return faq_data[best_idx]["answer"]
    return "Sorry, I don't understand that question."

Deployment on Render

Render makes deployment straightforward:

1. Connect your GitHub repo
2. Set build command: pip install -r requirements.txt
3. Set start command: gunicorn app:app
4. Done!

The free tier spins down after inactivity, so the first request after idle takes ~30 seconds. For production, you'd want a paid plan.

Future Improvements

• Add conversation context (multi-turn dialogue)
• Train on custom data with Word2Vec embeddings
• Add a web admin panel for managing FAQ entries

Live demo: faq-chatbot.onrender.com Source: GitHub