Skip to main content

Use cases and extensions

You have built a powerful semantic search system using Chromia’s Vector Database Extension.
Your movie embeddings are stored on-chain, and your Python client performs real-time semantic queries against them.

Now, let’s explore how to take this foundation further.

Real-world applications​

This vector search pipeline works seamlessly with any type of content, not just movies.

You can integrate:

  • Product descriptions for semantic ecommerce search
  • News articles for on-chain curation and retrieval
  • Support tickets for automated help agents
  • User posts or comments for recommendation engines
  • Knowledge bases for RAG-powered chatbots

All of these are backed by Chromia’s decentralized data layer and are searchable via vector embeddings.

Retrieval-augmented generation (RAG)​

The typical RAG workflow includes the following steps:

  1. A user submits a query.
  2. The system embeds the query.
  3. A vector database returns relevant results.
  4. A language model uses the results for context.

You have already implemented steps 1–3 on Chromia.

Now, complete the loop by connecting your semantic search results to a language model — for instance, in a chat interface.
The model can receive relevant matches as context and generate responses based on them.

The GOAT SDK course showcases a chat agent that interacts with Chromia, including tools, queries, and blockchain calls.

You can adopt a similar approach by:

  • Embedding the user prompt
  • Executing a vector query using the Python client
  • Feeding the result back into the agent as contextual input

This enhances your agent by making it retrieval-augmented, powered by fully on-chain data and semantic understanding.

🎉 Congratulations​

You have successfully built a decentralized semantic search system on Chromia.

From transforming raw text into on-chain vectors to enabling real-time querying, you now possess a complete and extensible pipeline, ready for real-world applications.

What’s next?​

With your pipeline established, you are prepared to:

  • Explore different models or embedding strategies
  • Add a chat or web interface
  • Combine with other on-chain logic or user actions
  • Apply the same semantic search structure across various domains

You’ve already laid the groundwork — everything else is an exciting extension!