Today’s AI models seem capable of practically anything. They play games autonomously, draw up architectural plans, and solve scientific problems.
But how do you build such powerful agents yourself?
Ailoy lets you build agents like these on your own machine, without needing to understand the complexities of AI theory.
Here are a few interesting demos of what you can do with Ailoy. You can try these examples yourself.
Image Editing (with SAM3)
CAD Modeling (with CadQuery)
Gameplay (OpenTTD)
Running Laya(a JEV-like decision model)
How Does It Work?
Most agent development frameworks focus on how to connect an LLM to tools (e.g. MCP). Ailoy can do that too, but it also offers a way to make agents far more powerful: give the agent a computer of its own.
This lets you build agents that do more than call predefined tools: they can install and use software, create their own scripts, and operate in a general-purpose computing environment (without terraforming the host system).
To make this possible, Ailoy gives each agent a virtual machine of its own without requiring you to install a separate VM daemon.
Regardless of your host OS, the agent can do the following inside its isolated Linux VM:
- Gather context for the task from accessible files and network resources.
- Write and run scripts to perform tasks and verify the results.
- Run ML models using the host’s GPU.
Your agent now has the superpower to tackle virtually any task. Just tell it what you need, and let it work out the rest.
Quickstart
See Ailoy's Quickstart for more details.
Links
Github: https://github.com/brekkylab/ailoy
Docs: https://brekkylab.github.io/ailoy/
Top comments (0)