The Open Standard for AI Blending Experiments
A structured format to record, share, and reproduce AI blending experiments. Fork others' work, track what works — and what doesn't.
Blend
Combine multiple AI models with different strategies — routing, cascading, ensemble, and merging — to build systems greater than the sum of their parts.
Benchmark
Track experiment performance with F1, precision, recall, latency, and cost metrics. Compare against baselines and identify what actually moves the needle.
Share
Fork experiments, discuss results, and collaborate with the community. Every experiment is versioned, reproducible, and open for iteration.
No experiments yet
Be the first to record an AI blending experiment. Document what you tried, what worked, and what didn't.
Create First Experiment