AI Experiments
Prototypes and probes at the edge of what AI can do.
Overview
Not a single product, but a laboratory. A place to run experiments, test new model capabilities, build custom agents, and discover interesting patterns.
Why I built it
The best way to understand rapid advancements in AI is hands-on experimentation across models, prompts, and tool-calling interfaces.
What I was exploring
Agent architectures, tool use, multimodal inputs, local model inference with Ollama, and AI-assisted development workflows.
How it works
Each experiment is self-contained. Some are CLI scripts, some web UIs, and others agent pipelines. The common thread is rapid iteration.
Technology
Current status
Active — A running collection of smaller AI experiments — agents, generative interfaces, automation scripts, and explorations with Gemini, OpenRouter, Ollama, and local models.