Demos - AO Labs | continuous AI training

Demos & Agent Reference Designs

Here are a few frontend apps (we❤️streamlit.io) to demonstrate applications of our Agents-as-a-Service API.

They are more conceptual than practical at this stage. If an application suits you, you can fork the whole App or the Arch of the Agent to customize around your data model.

All the code below is open source and available at our Archs repo with documentation available.
4 neurons: 3-to-1 channels <> 3-1 input-output neurons
Hello, World: a Clam-Level AGI
Our "Hello, World!" A 4 neuron system you can train with only input-side information to demonstrate our approach to AI.

Demo App

Fork Agent Arch
40 neurons: 3-to-1 channels <> 30-10 input-output neurons
Netbox Device Discovery: context-aware autocomplete for relational data
Hitting 85%+ accuracy on perdicting roles of new devices given as little as 70 examples of previous devices.

Demo App

Fork Agent Arch
788 neurons: 1-to-1 channels <> 784-4 input-output neurons
v0.1.0 MNIST Benchmark
Useful to get a feel for how the connections between neurons changes performance. Currently a script in our repo; demo app coming soon.

Fork Agent Arch
Arch Repo, Guides, & Tools

Archs & Apps Repo

Arch Config Guide

Demos & Agent Reference Designs