Science & methodology
Kysero turns broad exercise-science guidance into explicit, inspectable planning rules - then checks the result and keeps the uncertain parts editable.
Researchdoesnotproduceoneperfectplan.Itdefinesusefulprinciplesandranges;Kyseroappliesthemconsistentlytothegoal,person,scheduleandequipmentinfrontofit.
From input to decision
Open a row to see where the input acts and where Kysero deliberately stops making claims.
Strength, muscle gain, fat loss, general fitness, endurance and mobility do not share one prescription. The selected goal changes the weekly structure, eligible session roles, repetition or time ranges, rest, target effort and conditioning emphasis. These are goal-specific rules, not a claim that one routine is universally optimal.
The general principles
These principles influence the builder, but none of them identifies one universally optimal routine.
01
A plan should reflect the adaptation being trained. Kysero expresses this through goal-specific session roles and prescriptions instead of applying one set-and-rep template to strength, conditioning and mobility work.
02
Volume and frequency are considered together, then distributed across the days the user can actually train. Research supports useful targets and trends, but not one sharp biological minimum or maximum for every person.
03
Sets and repetitions are paired with rest and, where appropriate, repetitions in reserve. RIR is useful for communicating effort, but its accuracy varies, so it remains an editable target rather than a measurement Kysero pretends to observe.
04
Kysero keeps future prescriptions conservative until completed work, target effort, acceptable technique and recovery support a change. Requested deloads can be scheduled in eligible multi-week plans; the software does not infer that a person needs one.
Audit and verification
Software verification can show that the implemented rules behave consistently. It cannot convert those checks into proof of individual results.
01
Before a guided plan is returned, Kysero checks week and day counts, ordering, exercise references, complete prescriptions, session phases, exclusions, equipment conflicts, supported strict limitations and duration-estimate coverage. Structural failures reject generation; judgement-sensitive findings remain visible for review.
02
Automated suites exercise every supported goal and level, multiple split and frequency boundaries, limited-equipment cases, required and excluded exercise conflicts, limitation severities, multi-week behavior and previously repaired defects. These tests verify implemented behavior - they do not prove physical outcomes.
03
A separate audit corpus compares generated plans with published or official protocol arms across the supported goals. Differences are manually adjudicated as acceptable alternatives, missing information or genuine defects. Production generation never reads those protocol fixtures, so it is not reproducing memorized study plans.
04
Kysero has not run an independent clinical trial and does not guarantee strength, muscle gain, fat loss or injury safety. Qualified blinded review and adequate real-world completion and effort data remain confidence requirements, not evidence the product currently claims to have completed.
A useful boundary
Kysero cannot observe technique, pain, sleep, nutrition, measured aerobic intensity or how a future session will feel. Starting loads and RIR are planning targets, not measurements. Stop and seek appropriate professional advice when symptoms, injury, medical conditions or uncertainty make unsupervised training inappropriate.
Selected references
These papers inform the evidence layer. Kysero's exact thresholds, rankings and fallback behavior remain product rules unless stated otherwise.
2026 overview of reviews used for broad resistance-training guidance and the limits of prescription detail.
2023 systematic review and Bayesian network meta-analysis comparing combinations of load, sets and frequency.
2025 meta-regressions informing dose-response direction, diminishing returns and direct versus indirect set accounting.
2024 systematic review and meta-analysis supporting the view that split choice is mainly organizational when volume is equated.
2022 randomized trial showing that load and repetition progression can both be viable when progression follows observed performance.
2021 systematic review and meta-analysis used to support performance-responsive progression without claiming one universal method.
2022 study supporting RIR as an effort-prescription tool in its studied exercises and population, not as a perfectly observed value.
2024 systematic review and meta-analysis used to bound mobility guidance without turning one dose into a universal rule.
