SynthScope: Search, Visualize, Listen to Information

In this post, I will introduce you to SynthScope, one of my latest Google Gemini-based projects that enables a user to search the web and return search results as text, image, and audio simultaneously.

This post will give a high-level overview of the application. It will not discuss code implementation; just how to use the application for your daily information needs. Links to the application and codebase on GitHub are shared in this post.

What is SynthScope?
SynthScope is an LLM-powered tool that can be used to retrieve information from the web. Web search results powered by Google Search are returned as text and audio, and also converted into an image generation prompt, which is used to imagine the search result. You can also set SynthScope to translate the generated text and audio into any of 15 supported languages besides English, including Tamil, Thai, Japanese, and Arabic.

Features of SynthScope

• Text generation: Displays the search result in the preferred language text.
• Image generation: Displays the search result in the preferred image style out of 11 different styles.
• Audio generation: Speech capability reads out the search result.
• Language translation: Select the preferred language for the text and audio output.

How to Use SynthScope
Using SynthScope is very easy. Simply type in your search query, select the image style in which you want SynthScope to imagine the search result, select the preferred language from the language dropdown menu, and select the preferred voice of the reader from the voice dropdown menu.

Here is a diagram summary of how to use SynthScope:

With SynthScope, you can search for current information on the web and have it read out to you in your preferred language instead of scrolling to read text.

What Technologies Built SynthScope?
Here are the technologies that were used to build SynthScope:

• Python for writing the application logic.
• Google Gemini family of models for text generation, image generation, and text-to-speech (TTS).
• Gradio for frontend development.
• CSS for styling the frontend of the Gradio application.
• Hugging Face for deploying the application.

How to Access SynthScope
SynthScope is currently deployed on Hugging Face as a space. You can access it here.

Also, SynthScope is an open-source project, and that means that you can take a look at the code behind the application and even make contributions. You can access the code on GitHub.

I would appreciate your supporting the project with a Hugging Face like and a GitHub star, if possible :).

Limitation of Using SynthScope
The principal limitation of using SynthScope is that it is subject to the rate limits imposed on Google's Gemini models' free tier API. Here are the rate limits:

Text generation: Limited to 1500 requests per day
Image generation: Limited to 100 requests per day
Audio generation: Limited to 15 requests per day

Therefore, you may try to use SynthScope at a time when the daily quota for any of the above functionalities has been exhausted.

Conclusion
SynthScope is a creative way to search the internet for information. It is designed to be user-friendly and language dynamic, enabling users to read, visualize, and listen to information.