Voice Image Prompt Bot

Voice Image Prompt Bot v2 

First step in creating the voice interface is creating a new Telegram bot to communicate with. Follow these instructions here to create the bot and then connect it to Make: CLICK HERE

A brief overview of those instructions - 

● Go to the Bot Father 

● Send the message /newbot 

● Follow the prompts to name the bot 

● In Make, connect that bot through the “Watch for Updates” module 

Next, this scenario uses ModelsLab for A.I. image generation, so you need to sign up for an account with modelslab.com (CLICK HERE) 

This will give you the API key necessary to generate the AI art. 

How this part of the scenario works, the first ModelsLab HTTP module makes the API request to generate an image. Sometimes the image is generated immediately, sometimes it needs a few minutes. That is why there is a router. If the status comes back as “success” it grabs the image immediately. If it comes back as “processing” it waits 5 minutes then grabs the image. 

In the HTTP modules you MUST input your own API key.

Also, you’ll want to choose a public model.

Just copy/paste the desired model ID into the JSON in your http module. Beyond this, you’ll need to connect your OpenAI account to make the Whisper API work. Everything else is set up already. Enjoy your voice-to-image scenario. 

***IMPORTANT NOTE. This will not work with “voice notes” in Telegram. For some reason Telegram converts those into a file that the GPT Whisper API doesn’t recognize. You have to record a video and it will read the audio from that.