
🤖👁️ The AI Little Eye 🤫 🧪 I experiment with AI 📚 I spread AI knowledge 🛠️ I build AI products Join my creative frontier! 🚀 Created by @Juanky_Soriano
Joined March 2023
- Tweets370
- Following264
- Followers187
- Likes262
I am loving playing the mini game I did as a pet project during the weekend. It’s so much fun!
In case someone wants to play, it’s open source
github.com/juankysoriano/gpt…
1
49
Every now and then I like to do experiments with AI
Today I am presenting "GPT-Crimes"
A Telegram bot to play a conversational adventure in the style of those games in the early 80's
As always open-source
github.com/juankysoriano/gpt…
Enjoy!
@OpenAI @langchain #truecrime
Compose a song with AI in minutes! "🎼🤖
In a few steps, create an entire song using AI - it's all in the mix! 🎹🔥
#musicgen #audiocraft #gpt
1
181
Introducing "GPT-Producer"
Bring your Digital Piano, your @OpenAI API key, and your #GPU and become an artist!
#musicgen #audiocraft #gpt
github.com/juankysoriano/gpt…
Thanks @langchain @MetaAI for the tools that made this possible!
Retro-futuristic prompt alert! 🚨
This is from the new Podmesh site! Show me your interpretations! 🌌🎧 #Midjourney #AI
Prompt in ALT
1
188
The UI uses the @streamlit 'file_uploader' to upload a tmp file (bc. whisper wants a file path). An 'expander' with a 'text' elements are used to show the transcript. Finally, the summary is displayed. The 'spinner' and 'success' elements are used for loading indicator and result
1
93
The 'summarize' method takes the timestamp'd transcript text, splits it in chunks (to avoid token limit issues) and uses the @langchain 'load_summarize_chain' to summarize the transcript into a few sentences.
python.langchain.com/en/late…
1
67
The 'transcribe' function transcribes the audio file into text using @OpenAI Whisper and returns a string with the generated transcript with timestamps.
github.com/openai/whisper
1
56
Of course, don't forget to install dependencies and to use your own #OpenAI API key. Running the code generates a Streamlit web app that looks like this. It looks pretty good, if you ask me ✨! Keep reading for a code walkthrough.
1
25
Now, let me show you the code. Let's build a simple prototype. Imagine a tool that turns a podcast or speech audio file into a short summary. Guess what? It's possible in just 30 lines of code! Here's the full code, dependencies and command to start the app.
1
25
🤖🎉 Once done, a bot pings us on Slack to celebrate the successful completion of the workflow 🥳 Voila! Customized daily podcasts, thanks to automation and AI, with a touch of human finesse! 🎙️🧡
1
137
🌙 Workflow 3 - end of the day. We check our user list & ensure each gets a podcast on their interests they haven't heard before 🎯 Fresh content coming your way! 💌
1
96
🕛 Workflow 2 kicks in after noon - time for Text-to-Speech! Reviewed scripts are sent to @elevenlabs to create voices - the results are great!, in post-production they are combined with a prerecorded opening & ending 🎧
1
93
Step 5: We search for books related to the topics so we can make reading recommendations 🧡📚
Step 6: A Slack notification tells us the scripts are ready for human review ✅
1
103
















