02 · Epistemic Brewing
Every batch
is a question.
Mead, measured like a lab. Two research tracks at opposite ends of the alcohol spectrum, and a CSV log for every batch in between.
The style
Hydromel
A low-alcohol mead, roughly 3.5–7.5% ABV, finished dry and carbonated. Think of it as the session beer of honey wine: light, fizzy, drinkable in an afternoon rather than aged for a year.
Low gravity
Start around 1.030–1.055 instead of the 1.100+ of a standard mead. Less honey, faster ferment, drier finish.
Carbonated
Finished with a priming dose or force-carbonated in a keg. The fizz carries aroma that a still mead hides.
Quick turnaround
Days to weeks, not months. Which is exactly why it's the right style for controlled, repeatable experiments.
The research
Two ends of the spectrum.
Most of the batches here serve one of two long-running questions. Everything else is an excuse to try a strange ingredient.
Fastest clean hydromel
How quickly can a hydromel go from pitch to carbonated glass without off-flavors, if every factor is controlled?
- Yeast strain and pitch rate
- Temperature profile, held by a heat band and logged
- Staggered nutrient additions, dosed on schedule
- Oxygen at pitch, light exclusion, pH management
- Carbonation method and conditioning time
Highest ABV we can reach
Can we push a mead past the published alcohol tolerance of its yeast, and keep the yeast alive and working to get there?
- Step-feeding honey to manage osmotic stress
- Nutrient and temperature strategies for late-stage viability
- A yeast-longevity R&D tool, designed and built in-house (in development)
- Measuring where each strain actually stalls versus where the datasheet says
Batches
Every batch gets a page: the question, the recipe, the readings, the answer. Logs download as CSV.
Loading batches…
About the data
Each batch log is a plain CSV: timestamp, day, specific gravity, temperature, pH and a note column for events like nutrient additions. The index CSV summarizes every batch in one table.
Per-batch logs are open to download. As the research tracks mature, the deeper technical datasets, such as sensor-level logs, the yeast-longevity rig, and full process parameters, may move to a subscriber tier to support the work.