Alternative to Strava: a running log with ChatGPT

Set up a running log for me in databasedmcp, in one go.

1. Create a database called running_log, described as "Running log: every run, with distance, time and how it felt".
2. Create two tables:
   - shoes: id, name (required), started_on (date), retired_on (NULL while I still run in them).
   - runs: id, ran_on (date, defaults to today), kind (easy, long, tempo, intervals or race), distance_km (required), duration_s (required, total time in seconds), avg_hr, elevation_m, route, shoe_id (references shoes), felt (a few words on how it went), notes.
3. Write the database a guide that says what each table is for, that pace is worked out from duration_s and distance_km rather than stored, and the queries for "km per week for the last 8 weeks", "average pace by kind of run" and "km on each pair of shoes".

From then on: when I tell you about a run, or paste a summary from my watch app, add it to runs with the shoes I'm wearing. When I get new shoes, add them; when I retire a pair, set retired_on. Show me the tables when you're finished.

CREATE TABLE shoes (
  id INTEGER PRIMARY KEY,
  name TEXT NOT NULL,
  started_on TEXT,
  retired_on TEXT
);

CREATE TABLE runs (
  id INTEGER PRIMARY KEY,
  ran_on TEXT NOT NULL DEFAULT (date('now')),
  kind TEXT CHECK (kind IN ('easy', 'long', 'tempo', 'intervals', 'race')),
  distance_km REAL NOT NULL,
  duration_s INTEGER NOT NULL,
  avg_hr INTEGER,
  elevation_m INTEGER,
  route TEXT,
  shoe_id INTEGER REFERENCES shoes(id),
  felt TEXT,
  notes TEXT
);

# running_log: runs and shoes

- **runs**: one row per run. `duration_s` is total time in seconds. kind: easy, long, tempo, intervals or race. `felt` is a few words in my words.
- **shoes**: pairs I run in. `retired_on` stays NULL while I still use them.

Pace isn't stored: it's `duration_s / distance_km` seconds per km. Show it as min:sec.

Per week: `SELECT date(ran_on, '-6 days', 'weekday 1') AS week, count(*) AS runs, round(sum(distance_km), 1) AS km FROM runs WHERE ran_on >= date('now', '-56 days') GROUP BY week ORDER BY week`

Shoe mileage: `SELECT s.name, round(coalesce(sum(r.distance_km), 0), 1) AS km FROM shoes s LEFT JOIN runs r ON r.shoe_id = s.id WHERE s.retired_on IS NULL GROUP BY s.id ORDER BY km DESC`

INSERT INTO shoes (name, started_on) VALUES (?, ?) RETURNING id
-- params: ["Trailfox 3","2026-09-01"]

INSERT INTO runs (kind, distance_km, duration_s, route, shoe_id, felt) VALUES (?, ?, ?, ?, ?, ?)
-- params: ["easy",8.2,2790,"Lake loop",1,"Legs felt fresh"]

INSERT INTO runs (ran_on, kind, distance_km, duration_s, avg_hr, elevation_m, shoe_id, felt) VALUES (date('now', 'weekday 0', '-7 days'), ?, ?, ?, ?, ?, ?, ?)
-- params: ["long",18.04,6135,148,210,1,"Heavy in the last 3 km"]

INSERT INTO runs (ran_on, kind, distance_km, duration_s, shoe_id, notes) VALUES (date('now', 'weekday 4', '-7 days'), ?, ?, ?, ?, ?)
-- params: ["tempo",10,3130,1,"2 km warm-up, 5 km at tempo in 22:40, 3 km cool-down"]

SELECT date(ran_on, '-6 days', 'weekday 1') AS week, count(*) AS runs, round(sum(distance_km), 1) AS km FROM runs WHERE ran_on >= date('now', '-56 days') GROUP BY week ORDER BY week

SELECT kind, runs, km, printf('%d:%02d', pace / 60, pace % 60) AS pace_per_km FROM (SELECT kind, count(*) AS runs, round(sum(distance_km), 1) AS km, CAST(sum(duration_s) / sum(distance_km) AS INTEGER) AS pace FROM runs GROUP BY kind) ORDER BY km DESC

SELECT s.name, round(coalesce(sum(r.distance_km), 0), 1) AS km FROM shoes s LEFT JOIN runs r ON r.shoe_id = s.id WHERE s.retired_on IS NULL GROUP BY s.id ORDER BY km DESC