Home - davidmarsoni/llog GitHub Wiki

Cover page

:question: What is Llog?

Llog is a Project developed during a HES-SO Valais-Wallis 63-51 Emerging Technologies course. It is a custom chatbot that is connected to private data and public data to enhance AI responses in relation to specific subjects. This chatbot will also be able to reason through a chain of thought to provide precise and better structured answers.

:dart: What Is Llog For?

Llog is a project that aims to create a chatbot that can provide more accurate and relevant answers to user queries by leveraging private data and public data. The private data will mostly be a Notion integration that contains notes about the last 3 years of the HES-SO Valais-Wallis courses. The public data will be mostly internet or pre-trained data. We will also add some other data sources such as PDF or text files.

:wrench: What Problem Does Llog Solve?

Llog aims to solve the problem of providing accurate and relevant answers to user-specific queries by leveraging private and public data. This will avoid the problem of hallucination or generating false information. Having a chatbot that gives relevant and accurate answers to user queries is a big problem in the field of AI. E.g., if the chatbot gives false information about medical or legal information, it can have a big impact on the user's judgment or perspective on the matter.

:rocket: Open The Llog App

The Llog app is currently hosted on Google Cloud Run. You can access it by clicking on the button below:

[!NOTE] The production server takes about 30 seconds to start. This is because the project is stored on a serverless instance inside Google Cloud. This means that the server is not always running and it takes some time to start the server when you open the app. This is normal behavior for Google Cloud Run. If you want to have a faster response time, you can deploy the project on your local machine or on a virtual machine that is always running.

[!NOTE] The general app can experience a lot of slowdowns if multiple users are using it at the same time. This was done to avoid heavy costs on the Google Cloud Run service. To learn more about the costs of the project, you can check the Costs page.