# YazSes v2.9.0-dev.4 — Wave M continues (developer preview)

**Pre-release · not published** (publish workflows are pinned to `v0.x`/`v1.x`; v2 dev tags do
not ship). **v1.4.1 remains the stable release.** Every feature is **OFF by default**.

Continues **Wave M** (`design/vision/v2-research/15-wave-m.md`) with an output-formatting + a
knowledge-capture core:

## New since dev.3 (all off by default)

- **Screenplay Auto-Format** (ADR-v2-110, `[screenplay]`) — formats dictated scenes and dialogue as
  Fountain screenplay markup: "scene: interior coffee shop, day" → `INT. COFFEE SHOP - DAY`, "Maya
  (character) …" → an uppercased cue + dialogue, "transition: cut to" → `CUT TO:`, with smart curly
  quotes. `to_fountain` + `smart_quote_dialogue`. Anchor: Fountain (fountain.io).
- **Spoken Spaced-Repetition Capture** (ADR-v2-112, `[srscap]`) — say "remember that the capital of
  France is Paris" mid-dictation and it makes a local Anki cloze card ("The capital of France is
  {{c1::Paris}}"), scheduled by SM-2. `detect_fact` + `to_cloze` + `sm2_schedule`. Study material
  captured in the flow of work, no cloud. Anchor: Anki/SM-2.

`yazses features` now lists **123 capabilities**.

## Wave M remaining
Diagrams-as-Code · Interruptible Read-Back Proofreading.

## Quality
1399 tests green across Linux × macOS × Windows on Python 3.11 and 3.12. Both new cores are
unit-tested to 100% with **no new dependency** (pure regex + SM-2 arithmetic; deck export is a thin
downstream writer). The default install and the v1 dictation path are unchanged.