> For AI agents: the complete documentation index is available at https://open-spaced-repetition.github.io/ts-fsrs/llms.txt, the full documentation bundle is available at https://open-spaced-repetition.github.io/ts-fsrs/llms-full.txt. # DefaultScheduler `DefaultScheduler` is an async factory that returns a ready scheduler instance. It creates a model (FSRS-7 by default) with `dateChrono` and the standard scheduling policies, so the only thing you configure is its options. Internally it is the same definition-and-instance machinery as a custom `defineScheduler` — the preset is just a convenient assembly. ## Options When upgrading from FSRS-3\~6 to FSRS-7, omit `weights` to use FSRS-7 defaults or supply weights trained for FSRS-7. Replaying history rebuilds card state; it does not convert legacy weights. FSRS-7 cards include `stabilityFast` as well as `stability` and `difficulty`. Rebuild legacy FSRS-3\~6 memory states by replaying review history with `Reschedule`; parameter migration does not migrate card state. Do not seed an evolved fast trace with `0.8 * stability`. The preset enables fractional days for FSRS-7, and `enableShortTerm: false` only disables explicit learning steps, not the model's fast trace. For cost-adaptive retention, create a custom definition with `schedulerCostAdrMiddleware` in place of `schedulerDesiredRetentionMiddleware`. Cost ADR is not a `DefaultScheduler` option. `scheduler.nextInterval(memoryState, desiredRetention, { card, grade, elapsedDays })` queries the interval for an already computed post-rating state. It runs `handlers.nextInterval` without advancing the card or chrono; put shared interval policies in that hook. | Option | Default | Purpose | | -------------------- | --------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | `version` | `'FSRS-7'` | Selects the FSRS model: `'FSRS-3'`, `'FSRS-4'`, `'FSRS-4.5'`, `'FSRS-5'`, `'FSRS-6'`, or `'FSRS-7'`. FSRS-7 cannot migrate legacy FSRS-3\~6 weights. | | `weights` | preset weights | Optional trained weights; omit to use the preset's defaults for the selected version. FSRS-7 requires exactly 34 FSRS-7 weights; supplying legacy weights throws during scheduler construction. | | `desiredRetention` | `0.9` | Target retention applied before interval calculation. | | `enableShortTerm` | `true` | Enables learning and relearning steps. | | `learningSteps` | `['1m', '10m']` | Learning steps applied when `enableShortTerm` is on. | | `relearningSteps` | `['10m']` | Relearning steps applied when `enableShortTerm` is on. | | `graduatingInterval` | `1` | Shortest interval, in days, for a card in `Review`. A card which has a shorter interval after a review will stay in `Learning`/`Relearning`. When steps are configured, finishing them and rating a `Review` card `Hard`, `Good`, or `Easy` will result in at least this interval. | | `allowModelOverride` | FSRS-7: `true`; other versions: `false` | Uses the model interval instead of the current step when the model interval is longer. Again is capped at `graduatingInterval`; Hard and Good graduate early once the interval reaches `graduatingInterval`. | | `enableFuzz` | `false` | Spreads intervals for variety when enabled. | | `maximumInterval` | `36500` | Upper bound on the interval in days. | | `clearStatsOnForget` | `true` | Whether `forget` clears `reps` and `lapses`. | `DefaultScheduler` always uses `dateChrono` and applies the standard policies in the order described on the [Middleware](https://open-spaced-repetition.github.io/ts-fsrs/guide/middleware/built-in.md) page: desired retention, fuzzing, statistics, scheduled days, learning steps, maximum interval, then monotonic interval. Choosing `version` changes the model and its parameters, not this policy set. ```ts twoslash file="../../../snippets/run-code/default-scheduler.ts" import { DefaultScheduler, Rating } from 'ts-fsrs' const now = new Date('2026-01-01T00:00:00.000Z') const scheduler = await DefaultScheduler({ version: 'FSRS-7', desiredRetention: 0.9, enableShortTerm: true, learningSteps: ['1m', '10m'], relearningSteps: ['10m'], enableFuzz: false, maximumInterval: 36_500, }) const card = scheduler.newCard({ cardId: 'quick-start', now }) const result = scheduler.review({ card, grade: Rating.Good, now }) console.log( JSON.stringify( { state: result.card.state, dueAt: result.card.dueAt, stability: result.card.stability, stabilityFast: result.card.stabilityFast, difficulty: result.card.difficulty, rating: result.revlog.rating, }, null, 2 ) ) ``` ▶ Run code The factory is async, so `await DefaultScheduler(...)` before using the returned instance. The example makes the common defaults explicit: `version`, `desiredRetention`, learning and relearning steps, fuzzing, and `maximumInterval`. Omit `weights` to use the preset for the selected FSRS version. ## Preset or custom definition Reach for `DefaultScheduler` when the default FSRS model, `dateChrono`, and the standard policy set fit your product — it is the fastest correct start. Reach for `defineScheduler` when you need a different model or chronology, add or remove policies, or expose a configuration you validate per user. Both end in the same scheduler instance API (`newCard`, `review`, `rollback`, `preview`, and `forget`). See [Scheduler overview](https://open-spaced-repetition.github.io/ts-fsrs/guide/scheduler/index.md) for the definition-and-instance model.