We all know it. The Great Bull Run is about to kick into full gear. And that means it's time to sell all your belongings for sats.

OK that's hyperbole but you should still probably be trying to stack sats.

CASCDR has a new demo that can help.

Using the power of AI Vision and LLMs to reason about the image of an object, we can get advice about how much we ought to sell an item for in sats.

Stay Humble, Get an Appraisal, & Stack Sats

We recently created an endpoint that leverages GPT 4's vision capabilities which allows the world renowned LLM to consume an image, then reason about it. We then packaged that into a simple example that has immediate utility. You can use this service anonymously by simply paying with sats through your Alby extension.

What does that all mean for you?

Snap a photo. Click the Run Button. And get an appraisal of the item in satoshis with the help of AI.

Let's run through a quick example. Start by going to CASCDR, enter the main application, then go over to the Image2Text Playground Tab. Then we're ready:

Step 1: Snap a Photo of the Object to be Sold

Step 2: Upload the Image and Hit "Run"

Step 3: Wait a Few Seconds & Get Your Appraisal!

And there you have it. In a few seconds, you can accomplish a critical task in your Sats Stacking Journey by appraising the value of your stuff and getting it ready to sell on Craigslist or https://satscrap.com.

This simple example shows what's becoming possible with AI. We could easily foresee an application like this being wrapped inside of an e-commerce application that automatically sets pricing for you to ensure you make the deal & keep the sats flowing.

Other Examples to Drop

We didn't stop there for this demo. We added a couple more:

• Shopping Assistant: Need to buy something but don't know where to look or what keywords to use? Use the Shopping Assistant example and get a quick link to

• Custom - the magic is all in the follow up prompt. GPT4 can give you a pretty wide range of answers & utility depending on how you engineer a prompt. Hit the Image2Text Playground, press "Custom" and give your own follow up prompt a whirl.

CASCDR helps me purchase a tool I need even if though too smooth brained to know what it's called...

Summary of Available Services

On this CASCDR/Data Buffets project we have been hard at work building the primitives needed to get the most out of the AI Revolution while still respecting the cyberpunk values of privacy & freedom from surveillance.

So far we have added:

• 💬 General Purpose LLM (GPT)

• 📷 AI Vision

• 🗣️ AI Voice to Text Transcriptions

• 🎥Youtube Video Extraction

And put them in simple easy to use examples the demonstrate not only their value but what's possible. We hope this can spur the creativity & imagination of the community to create powerful applications.

Roadmap of What's Next

Now that we have sufficient services available, we will be shifting our focus to making a flexible application that allows users to create arbitrary workflow apps!

To do that, we will be doing an overhaul to the NIP105 spec that allows users to search services on the basis of their input/output schemas. We do this by compressing that information into a simple SHA256 hash and letting the clients figure out "what's compatible with what".

Finally, we will add this capability to CASCDR to demonstrate & allow our users to build their own workflows based on whatever they want to accomplish!