Embedchain

Embedchain is an open-source Retrieval-Augmented Generation (RAG) framework used to create AI applications that answer questions from custom documents and other data sources. This deployment uses local Ollama models and does not require an external API key.

Version: 0.1.128

The solution includes the Ollama runtime with the Llama 3.2 1B language model and the All-MiniLM embedding model. It is suitable for document question answering, knowledge-base search, local AI experimentation, and RAG application development.

Features of Embedchain:

• Builds AI-powered RAG applications using custom data.
• Supports text files, documents, websites, and other data sources.
• Uses local Ollama models without requiring an API key.
• Generates embeddings and stores searchable document information.
• Provides a simple Python interface for adding and querying data.
• Suitable for private knowledge bases and document question-answering applications.

Usage instructions for Embedchain:

Connect to the virtual machine using SSH and switch to the root user:

$ sudo su

Open the Embedchain installation directory:

$ cd /opt/embedchain

Activate the Python virtual environment:

$ source .venv/bin/activate

Check the installed Embedchain version:

$ uv pip show embedchain

Check the Ollama service:

$ systemctl status ollama --no-pager

View the installed local AI models:

$ ollama list

Run the Embedchain application:

$ python app.py

Configuration file path:

/opt/embedchain/config.yaml

Sample data file path:

/opt/embedchain/data/company.txt

Application file path:

/opt/embedchain/app.py

Embedchain is a CLI-based application. It does not provide a web dashboard or browser URL in this deployment.

Ollama runs locally at:

http://127.0.0.1:11434

Disclaimer: Embedchain is provided “as is” under applicable open-source licenses. Users are responsible for configuring data sources, securing sensitive information, validating AI-generated responses, and allocating sufficient compute resources. This solution is best suited for RAG development, local AI experimentation, document search, and knowledge-base question-answering use cases.