Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/jaimegayo/KERNDOCUMENTATION/llms.txt

Use this file to discover all available pages before exploring further.

Every time you finish a training session in KERN, the app compiles a complete snapshot of your performance — every exercise, every set, every weight — and sends it to the API. That record is what powers the progression hints you see the next time you open the same routine.

​
The WorkoutSession model

Each completed session is persisted as a WorkoutSession row in the database. The table schema is defined in models.py:

Session identity

id (primary key), user_id (foreign key to users), routine_name (string copy of the routine name at the time of the session), created_at (server timestamp).

Performance data

duration_seconds (integer), total_volume (float — sum of kg × reps), steps (integer — from the step counter service), data_json (JSON — full exercise and series snapshot).
class WorkoutSession(Base):
    __tablename__ = "workout_sessions"

    id              = Column(Integer, primary_key=True, index=True)
    user_id         = Column(Integer, ForeignKey("users.id"), nullable=False)
    routine_name    = Column(String(255), nullable=False)
    duration_seconds= Column(Integer, nullable=False)
    total_volume    = Column(Float, default=0.0)
    steps           = Column(Integer, default=0)
    created_at      = Column(DateTime(timezone=True), server_default=func.now())
    data_json       = Column(JSON, nullable=False)

​
The active workout screen

When you open a routine, KERN calls GET /routines/{routine_id}. The response includes historical data from your last session alongside the current exercise structure. Each series object in the response carries four fields relevant to progression:
FieldTypeDescription
anteriorstringHuman-readable hint, e.g. "60kg x 10". Displayed in the UI.
prev_kilosfloatNumeric weight from the last session. Pre-fills the input field.
prev_repsintNumeric rep count from the last session. Pre-fills the input field.
kilos / repsintAlways sent as 0 — you enter new values during the session.
The anterior hint gives you a quick reference target without forcing you to match it. Enter whatever weight and reps you completed — KERN records what actually happened, not a goal.

​
Volume calculation

Total volume is the sum of kilos × reps for every completed series in the session. The Android app (WorkoutActiveFragment) calculates this in real time via the ActiveExerciseAdapter.OnVolumenChangeListener callback and updates the header counter as you mark sets done. The final value is included in the POST /workouts/finish payload as total_volume.

​
Finishing a session

When you tap “Finish workout”, the app calls POST /workouts/finish with the WorkoutSessionCreate schema: 
{
  "routine_name": "Mi Rutina Full Body Inicial",
  "duration_seconds": 3420,
  "total_volume": 4850.0,
  "steps": 312,
  "data_json": [
    {
      "name": "Press de banca",
      "series": [
        { "numSerie": 1, "kilos": 60, "reps": 10 },
        { "numSerie": 2, "kilos": 60, "reps": 9 },
        { "numSerie": 3, "kilos": 55, "reps": 10 }
      ]
    },
    {
      "name": "Dominadas",
      "series": [
        { "numSerie": 1, "kilos": 0, "reps": 8 },
        { "numSerie": 2, "kilos": 0, "reps": 7 }
      ]
    }
  ]
}
Only series where isDone = true are included in data_json. Empty exercises (no completed series) are filtered out on the client before the request is sent.

​
Rest timer

After you mark a set as complete, the rest timer activates automatically. It is managed by RoutineViewModel and displayed via a dialog in WorkoutActiveFragment. You can adjust the duration in 15-second increments using the +15 and -15 buttons, or reset it entirely. An audio tone plays when time runs out.

​
Historical matching logic

When loading a routine, KERN matches series from your last session to the current routine’s series using two strategies:
1

Match by numSerie (primary)

The API builds a map of numSerie → series data from the last session. For each current series, it looks up the same numSerie value in that map. This is the preferred path because series numbers survive routine edits.
2

Match by index (fallback)

If no numSerie match is found — for example, with sessions recorded before the field was introduced — the API falls back to positional matching using the series’ array index.

Build docs developers (and LLMs) love

Get started for freeTalk to us