This product was not featured by Product Hunt yet. It will not be visible on their landing page and won't be ranked (cannot win product of the day regardless of upvotes).
CAFE — Compound-AI Factorial Evaluation
Stop guessing which AI config is better. Prove it.
CAFE treats every knob in your AI pipeline - retrieval, reranking, prompts, models, and tools - as an experimental factor. It runs factorial experiments, evaluates outputs using a configurable LLM (and optionally human reviewers), and applies mixed-effects models to determine: - Which techniques actually improve quality - How much each technique contributes - Whether the observed differences are statistically significant Open source and self-hostable.
Hi! As part of a research paper, we built CAFE to answer a question that kept coming up: when I tweaked my RAG or agent pipeline and the output improved, which change actually made the difference? Aggregate benchmarks and eyeballing a handful of outputs never really answered that - especially when LLMs are nondeterministic from run to run.
CAFE treats every knob in your pipeline (retrieval, reranking, context assembly, prompts, models, tools, etc.) as an experimental factor. It:
- Generates a full or fractional factorial design - every configuration combination worth testing
- Runs each configuration as a black box with replication (concurrent and resumable)
- Scores outputs using a configurable LLM judge and/or human raters
- Attributes performance differences using mixed-effects models matched to your rubric's scale
The result is a statistically grounded answer to questions like:
- Which techniques actually improve quality?
- How much does each technique contribute?
- What is the best-performing configuration?
- Are the observed differences real, or just noise?
CAFE also includes a cost–quality Pareto frontier and judge↔human agreement analysis using Krippendorff's α.
It's open source (Apache-2.0) and fully self-hostable. You can use it as a Python librar or a FastAPI + React web application. Nothing leaves your machine.
⭐ GitHub: https://github.com/fabian-lu/Cafe
🧪 Live demo: https://cafe-ai.de/demo
📚 Documentation: https://fabian-lu.github.io/Cafe
I'd love your feedback!!
finally a tool that says "statistical significance" and means it. love that you can swap the eval LLM and rerun everything without rebuilding the whole experiment.
Would love to see automatic cost and latency tracking baked into each experimental factor so you can weigh quality gains against compute spend, not just statistical significance. Right now it’s all about the output score, but in production the more expensive combo that wins by 2% isn’t always the right call.
About CAFE — Compound-AI Factorial Evaluation on Product Hunt
“Stop guessing which AI config is better. Prove it.”
CAFE — Compound-AI Factorial Evaluation was submitted on Product Hunt and earned 8 upvotes and 5 comments, placing #159 on the daily leaderboard. CAFE treats every knob in your AI pipeline - retrieval, reranking, prompts, models, and tools - as an experimental factor. It runs factorial experiments, evaluates outputs using a configurable LLM (and optionally human reviewers), and applies mixed-effects models to determine: - Which techniques actually improve quality - How much each technique contributes - Whether the observed differences are statistically significant Open source and self-hostable.
CAFE — Compound-AI Factorial Evaluation was featured in Open Source (68.6k followers), Developer Tools (515.9k followers), Artificial Intelligence (473.7k followers) and GitHub (41.3k followers) on Product Hunt. Together, these topics include over 221.6k products, making this a competitive space to launch in.
Who hunted CAFE — Compound-AI Factorial Evaluation?
CAFE — Compound-AI Factorial Evaluation was hunted by Fabian Lukassen. A “hunter” on Product Hunt is the community member who submits a product to the platform — uploading the images, the link, and tagging the makers behind it. Hunters typically write the first comment explaining why a product is worth attention, and their followers are notified the moment they post. Around 79% of featured launches on Product Hunt are self-hunted by their makers, but a well-known hunter still acts as a signal of quality to the rest of the community. See the full all-time top hunters leaderboard to discover who is shaping the Product Hunt ecosystem.
Want to see how CAFE — Compound-AI Factorial Evaluation stacked up against nearby launches in real time? Check out the live launch dashboard for upvote speed charts, proximity comparisons, and more analytics.
Fabian Lukassen