Quick Links

• Talk to page
• Getting Started
• Flowchart
• Screen Recording
• Features
• Contributing

Talk to page

• This project let you chat with a given url particularly research paper links.
• The poject aims to use the advancements of LLM while using the web pages as the context.
• With advance RAG (Retrieval Augument techniques) we can chat with the web pages and get the most relevant information.

Getting Started

Important

In the .env, file make sure to put your own Huggingface API token enable with Inference.

• Clone the repository from github.
• cd talk_to_page

Tip

Using Makefile

• RUN make start to start the agent for the first time
• RUN make run to restart the agent
• Project will be available at http://localhost:8501

Tip

Using Doker

• docker build -t talk-to-page:latest -f Dockerfile .
• docker run -p 8501:8501 talk-to-page:latest
• Project will be available at http://localhost:8501

Tip

Using pip

• pip install .
• CMD start -Project will be available at http://localhost:8501

Flowchart

Note

Searching the internet component is not release!! Please contact the developer for more information.

Screen Recording

🎥 Download the recording

Features

• Chat with web pages
• Get the most relevant information
• Use the web pages as the context
• Use the advancements of LLM
• In the Query provide a url, [there is no need of a specific format]

Contributing

• Fork the repository
• Create a new branch

Note

The project will use Haystack and webui to chat with the web pages and get the most relevant information.

Tip

The project development is pause * [ ].

Note

The project is open source and free to use. The project is a mid level RAG based AI Agent, which can be used to chat with web pages and get the most relevant information. For more advanced RAG based AI Agent, please contact the developer.