Experiments and their runs, with parameters, seeds and environment captured per run
Experiment tracking
Log runs, parameters, seeds, environment and outcomes, then reuse those records when you write the Methods section.
Start free
What it adds to the project
Concrete capabilities, kept close to the sources and decisions they support.
Code references linking a run to the commit that produced it
Results recorded next to the settings that generated them, not in a separate spreadsheet
Methods and settings flow into your manuscript instead of being reconstructed from memory
Shared with your lab, so a run is still legible after the person who ran it graduates
How it works
A short path from your material to something you can inspect, revise and reuse.
Create the experiment
One experiment per research question. It becomes the container for every run you try against it.
Log each run
Record parameters, seed, environment, the code reference and the outcome. A failed run is worth logging - it is what stops you repeating it.
Write from the record
When you draft the paper, the Methods and experimental-setup details come from what you logged, so the reported protocol matches the one you actually ran.
Common questions
- Is this a replacement for Weights & Biases or MLflow?
- Different job. Those instrument training loops and stream metrics. Labrix tracks experiments at the level a paper needs - settings, seeds, environment, outcome - and connects them to the manuscript. Many labs run both.
- Does it work for wet-lab work?
- Yes. Runs are not assumed to be code: a protocol, a reagent batch, a cohort or an instrument setting fits the same structure.
- Can my lab see my runs?
- If you share them. Team members can view and edit shared records; deleting stays with the owner.
Continue through the workflow
Try it with your own material
Labrix is free to start, with no credit card required for the basics.