We launched BuddhiAI on 5th November, 2015 as a browser based private AI. Since then, BuddhiAI has evolved a lot. With the release of Gemma 4, BuddhiAI is now powered with Gemma 4 E2B model. In this post, we are going to see, how we can use BuddhiAI.
First go to BuddhiAI app, and click “start” call to action button. This will launch the BuddhiAI app.
At first, if you are launching the app for the first time, you will see a message saying “model loading failed” with a button “Manage Models”. Click the button to launch the Model manage.
In the model manager page, you can see there are two models: Gemma 4 E2B and EmbeddingGemma 300m. Download both the models. This may take a while. You can leave it as it is and continue with your other work.
Once the model download is completed, click “New Chat” from the left side bar. Now you can see the chat UI is active.
Now you can chat with BuddhiAI (powered by Gemma 4 model) similar to how you chat in ChatGPT. By default, Thinking mode (Reasoning) is off. If you want, you can enable thinking mode by clicking the “Reasoning” toggle button in the left bottom of the chat input. Note: thinking will take some time to generate the final response.
Chat with your documents
The best part of BuddhiAI is, you can feed your documents to BuddhiAI and ground answers based on your documents. BuddhiAI has a simple RAG system implemented, where you can upload your documents and provide it as knowledge to the model. This RAG system runs completely within your browser and non of your data will be send away from your browser.
To feed your documents to BuddhiAI, click “Documents” from left sidebar, it will launch the Document Manager UI.
Click “Upload” tab, and upload your files. You can upload up to 5 files at a time and files must be less than 25 MB in size. (for now). The chunking and generating embedding process takes some time. Once the process is completed, navigate to a new chat window and start asking questions related to the documents you have uploaded and see how BuddhiAI answer your questions based on the information in the documents you have uploaded.
Also, checkout the video guide below: