Podchat.ai

Your own personal co-host for your podcast needs!

Comment

Inspiration

Podcasts have been rapidly growing over the past couple years, with there being over 450 million podcast listeners worldwide in 2023 alone. As podcasts continue to grow, I noticed that there was a place for AI to gather and summarize information for hosts whether before they go live or during. Additionally, as concerns about misinformation from podcasts rise, I found a need to make this tool to better ensure that hosts are more information about the topics they discuss, while allowing for extra practicality in sharing notes for the overall broadcast itself.

What it does

By running the command "flet run --web" in the "test/src" folder, you are greeted to an open chatroom that can be accessed by multiple users. Once inside the chatroom, users can talk to each other as they would on any messaging platform. On top of standard messaging, by typing the following commands, the user can get specific, AI assisted, info on whatever topic they are interested it, via Podchat's own Jenna.

  1. '!eli5' -> Have a desired topic be explained at the level of a 5-year-old. Great for beginners of any subject, encouraging intuitive learning.
  2. 'desc' -> Get a detailed description on any topic. Great for avid readers, giving more than enough facts & information for conversation and learning.
  3. '!sum' -> Summarize any topic the user inputs, concisely explaining any topic. Great for finding clarification in conversation & ending conversations.
  4. '!rel' -> Gives you a list of related topics and people on a subject. Great for brainstorming ideas & continuing the flow of conversation.
  5. '!wq -> Get a quick, 1-line description of any subject. Great for jogging your memory in those tip-of-the-tongue moments.

How we built it

I used python to build the entire project. Specifically, using the 'flet' library which is a python binding for flutter that lets you develop and build cross-platform apps using just python. I also used Google's Gemini API as the AI service for Jenna, the in-chat assistant. Giving the AI a real, common name also helped to humanize the AI and make interactions with it feel more conversational rather than transactional.

Challenges we ran into

At first, python was not the language I planned to use. The first couple hours I attended was rough as I bounced from language to language, framework to framework, just to end up using the classic python. After that, most of the challenges came from debugging issues with the flet library in terms of rendering and attempts to implement multiple pages inside our one app, which ultimately was not achived.

Accomplishments that we're proud of

Being able to find a practical tool that helped development time rather than hindered it. Additionally, I'm proud I was able to make it over the span of a day as a solo dev, while still attending workshops and making connections. I met a lot of friends and gained knowledgeable insights in fields like cybersecurity, AI Training, App Dev, and I'm happy that I took notes during those workshops as I will be going back to them soon.

What we learned

Balance priorities when building a project, especially on such short notice. As the deadline rings closer, bugs and issues you couldn't even conceive will come at the most inconvenient times. However, don't only work on your project, as you'll miss out on valuable experiences and lessons. Best to keep a balance of the both.

What's next for Podchat.ai

More development time. Would be fun to finish during the summer. Very useful.

Built With

Share this project:

Updates