Building a Simple Powerful Chatbot Using ChromaDB and Sentence-Transformers

Created an AI chatbot using ChromaDB and Sentence-Transformers. These tools made the whole process smooth and efficient, and I wanted to share my experience in hopes that it might help someone else starting out with similar projects.

Steps I Took to Build the Chatbot:
I used sentence-transformers to convert text data into embeddings and ChromaDB for vector storage.

so install these dependencies..

pip install -U sentence-transformers

pip install chromadb

Next

import chromadb
chroma_client = chromadb.Client()
collection = chroma_client.create_collection(name="data")

Don't forget to load your dataset

Using Sentence-Transformers, I generated embeddings for my dataset:

for index,text in data.iterrows():
   res=text['Answer']
   embedding = model.encode(res)
   collection.add(
       documents=[str(res)],
       embeddings=[embedding.tolist()],
       ids=[str(index)],
       metadatas=[{'text':res}]
   )

*Query and Response: *

The chatbot listens for user input, converts the query into an embedding, and uses ChromaDB to find the closest match in the stored embeddings. Here’s the basic query

def get_response(text):
    res=model.encode(text)
    result=collection.query(
        query_embeddings=[res.tolist()],
        n_results=1
    )
    return result

That's it ,If you pass any text relevant to your dataset it'll response for you which is has close distance

---

👇👇👇👇👇

I'm excited to introduce Terminal Troubleshooters — a dynamic Slack community for tech enthusiasts to learn, share, and grow together! Our mission is to build a collaborative space where you can:

✨ Share daily tech insights and experiences
💡 Discuss innovative ideas and industry trends
🗣️ Enhance communication skills through engaging conversations
🤝 Connect with like-minded professionals and expand your network

Community Link:

terminaltroub-egk3921.slack.com