Methodology · current product

How ULxInfinity chooses your next workout

A transparent explanation of the deterministic rules behind ULx∞ v0.8—including what its numbers mean, what they do not mean, and how to evaluate a recommendation.

1. Saved workouts, not a calendar script

The Action Audit reads the saved sessions inside a configurable rolling history window. Unsaved draft sets do not count as completed work. A past set contributes to the movement actions assigned to that exercise. ULx∞ estimates which actions are below, within, or above the target range configured by the user.

2. Weighted action coverage is an estimate

Not every muscle is equally involved in a compound lift. For planning, v0.8 uses the following default action-credit conventions. These weights help compare possible choices—they are not laboratory measurements of equivalent hypertrophy stimulus.

Incidental stabilization is not automatically treated as meaningful direct work. The exercise library has specific metadata and selected exceptions to the general weights; older saved effect snapshots remain authoritative so later catalog edits do not silently rewrite history.

3. Priorities change decisions, not history

Bodybuilding, Powerlifting, Powerbuilding, and General Strength Training Styles provide editable starting priorities. They all configure the same underlying generator; there is no separate hidden model for each style. Custom allows manual configuration. Bench, squat, and deadlift competition specificity is tracked separately from muscle-coverage needs. A lift-specific goal should not be confused with a broad hypertrophy set count.

4. Build a coherent session one exercise at a time

ULx∞ evaluates whether Upper or Lower has the more pressing useful deficit, considering your settings. As it selects an eligible exercise, it recalculates projected coverage before choosing another one. This reduces needless repetition and helps leave room for other deficient areas. Close substitutes are curated by movement purpose: a different angle or goal can still justify two exercises from the same broad muscle group.

Favorites serve as bounded preferences in close choices. Excluded exercises are hard restrictions for automatic generation. Composition checks prevent a single well-scoring muscle from automatically monopolizing the whole session, while still allowing extra work when need justifies it. Session length responds to useful remaining work, rather than promising a fixed number of exercises every time.

5. A concrete illustration: pulldowns and curls

Suppose a saved session contains two sets of a pulldown, with the movement's back action marked primary and biceps involvement marked secondary. In the simplified planning ledger, that could add 2.0 back-action credits and 1.0 biceps-action credit. If biceps still have a meaningful deficit, a curl may remain useful; if their estimated coverage is already adequate, another exercise may rank higher.

This is an illustrative calculation, not a screenshot of a particular generated session or a physiological claim. Actual recommendations depend on action metadata, your configured targets and priorities, other saved work, eligibility rules, and composition scoring.

6. What this release does not claim

Version 0.8 does not measure your recovery, automatically calculate personalized maintenance volume, infer muscle growth from one performance change, or provide future progression-engine functionality. It does not use a generative model to invent workouts. It also does not enforce universal “press plus fly” or “row plus pulldown” quotas, or award generic muscle-growth bonuses simply because a movement uses a machine.

What to do next

If you have not used the planner yet, follow the first-workout walkthrough. You can also send generator feedback, ideally explaining your settings, expected result, and the mismatch—without sending personal workout data you do not wish to share.

Try the free planner →