01
Design
Structure the hypothesis, protocol, assumptions, metrics, data contracts, and success criteria before the experiment runs.
Origami Research
The operating system for experimentation.
•Origami Research
The missing layer
Trustworthy experimentation needs a system. Hypotheses, protocols, runs, metrics, artifacts, and the next experiment—all connected.
Experiments scatter.
Protocols live in docs. Metrics live in dashboards. Code lives in repos. Decisions live in conversations.
Context disappears.
Teams can see the final result, but not always why an experiment was run, what changed, or what should happen next.
Learning doesn't compound.
Every experiment should become reusable evidence—searchable, reproducible, and ready to inform the next run.
From ideas to reusable experiments and evidence
Start from an idea, paper, repo, dataset, or baseline. Origami helps you design the experiment, run it reproducibly, evaluate the evidence, and preserve what your team learns.
02
Run
Execute code and evaluations in a controlled sandbox while capturing parameters, artifacts, logs, blockers, and environment details.
03
Evaluate
Connect results back to the original question. See what changed, what passed, what failed, and whether the evidence supports the claim.
04
Learn
Save reusable methods, loaders, metrics, reports, and decisions so each experiment compounds into institutional knowledge.
Join the waitlist
Private beta with a small group of researchers and teams.
About Us
We're a small technical team with deep familiarity about how good research gets done. We have worked in labs conducting cutting-edge research in machine learning, robotics, and computer systems, and we're building the missing operating system for research experiments.
Our bet: the next wave of scientific progress will come from systems that facilitate deep human-AI collaboration in designing, running, tracking, and learning from experiments.
