@koda-sl/baker-cli 0.432.0-dev.795979397 → 0.434.0-dev.42b55fa66

package/README.md

@@ -4777,7 +4777,7 @@ array as a literal list: `"reference": ["$ref:subject.sheet", "$ref:type.image"]
 |---|---|---|
 | `images` | image[] | `image/png \| image/jpeg \| image/webp` (provider-determined) |
 
-**Cost** — 5 credits estimate per image asked for (`n`, default 1), 20 on Gemini 3 Pro Image (as `image_aspect_adapt` quotes it) and 20 at quality `max` (a static ad's re-laid-out sizes); the backend charges the call's actual OpenRouter usage.
+**Cost** — 5 credits estimate per image asked for (`n`, default 1), 20 on Gemini 3 Pro Image (as `image_aspect_adapt` quotes it) and 20 at quality `max`; the backend charges the call's actual OpenRouter usage.
 
 **Several candidates from one call** — `n` (integer, default 1) asks one call for several images, returned together in `images`. Each image is billed as if it were its own call (`n: 3` costs three images); what it saves is requests and waiting. Only models that declare it take it (today `openai/gpt-image-2.5-sunburst`, up to 10); any other model refuses it before anything is billed. Wire the whole array into `image_select` to keep the best.
 
@@ -5997,9 +5997,9 @@ Turn a source/inspiration image into a **runnable, self-validated static-ad canv
 
 It then scaffolds a canvas on which **the image model designs the finished ad**, holding only what `prompt.json` holds; no element is mandatory:
 
-- Up to **three variants** (`--takes`, default 3) render on the model the router picks for the concept (today `openai/gpt-image-2.5-sunburst` at quality `max`; `--gen-model` overrides), each a finished ad: the picture and every word of the copy. Each variant is **its own request** (tagged `static_variant_call`), run in parallel on its own direction (the scene as written; another picture of the same message, by the scene's look; another line leading, named by number so no word changes), and a `collect` tagged `static_variants` gathers them, on every model: one request with `n` costs the same (`n` is billed per picture) but gave fewer distinct options, so the graph never uses it. The brief quotes every line of `copy` verbatim and asks for each once; it is compiled again from `prompt.json` on every `validate` and `run`, so an edit re-keys exactly the renders it touches. The brand's **typeface** is shown to the model as a `font_specimen` set on the `designed` lines: a file in `src/brand/fonts/` (via `src/brand/tokens.json` when present), else the heading family `tokens.json` or `BRAND.md` names, fetched as a web font file; an `in_scene` line keeps its own lettering only on the real packaging, signage or screen its `on` names (handwriting on post-its, whiteboards or chalkboards, and a line naming no surface, are set as designed type). Each line's `role` sets its size: a price or stat (else the headline) leads, the `cta` is a large solid button, fine print is smallest. **The logo is optional**: the real file in `src/brand/logos/` goes to the model as a reference to reproduce once, at least a fifth of the frame wide, where it contrasts or on a small brand-colour plate, only when `prompt.json` has `"logo": true` (an older blueprint: an `advertiser_brand` entry in `brands_logos`) and the file exists; no logo is never an error, never a slot to fill and never a stand-in. Setting it false drops the logo on the next run; adding one to a canvas scaffolded without the file needs the file in `src/brand/logos/` and a re-scaffold. `prompt.json` also carries a **`brand`** block (ground, ink, accent, CTA and palette, from `src/brand/tokens.json` roles or `BRAND.md`); a colour outside it never reaches the model. The brief describes a complete ad at the master's shape and asks only for its edge margin (see below). `baker canvas set-prompt` refuses these nodes and names `prompt.json`.
+- Up to **three variants** (`--takes`, default 3) render on the model the router picks for the concept (today `openai/gpt-image-2.5-sunburst` at quality `xhigh`, about 45 s a picture against 80 s at `max`; `--gen-model` overrides), each a finished ad: the picture and every word of the copy. Each variant is **its own request** (tagged `static_variant_call`), run in parallel on its own direction (the scene as written; another picture of the same message, by the scene's look; another line leading, named by number so no word changes), and a `collect` tagged `static_variants` gathers them, on every model: one request with `n` costs the same (`n` is billed per picture) but gave fewer distinct options, so the graph never uses it. The brief quotes every line of `copy` verbatim and asks for each once; it is compiled again from `prompt.json` on every `validate` and `run`, so an edit re-keys exactly the renders it touches. The brand's **typeface** is shown to the model as a `font_specimen` set on the `designed` lines: a file in `src/brand/fonts/` (via `src/brand/tokens.json` when present), else the heading family `tokens.json` or `BRAND.md` names, fetched as a web font file; an `in_scene` line keeps its own lettering only on the real packaging, signage or screen its `on` names (handwriting on post-its, whiteboards or chalkboards, and a line naming no surface, are set as designed type). Each line's `role` sets its size: a price or stat (else the headline) leads, the `cta` is a large solid button, fine print is smallest. **The logo is optional**: the real file in `src/brand/logos/` goes to the model as a reference to reproduce once, at least a fifth of the frame wide, where it contrasts or on a small brand-colour plate, only when `prompt.json` has `"logo": true` (an older blueprint: an `advertiser_brand` entry in `brands_logos`) and the file exists; no logo is never an error, never a slot to fill and never a stand-in. Setting it false drops the logo on the next run; adding one to a canvas scaffolded without the file needs the file in `src/brand/logos/` and a re-scaffold. `prompt.json` also carries a **`brand`** block (ground, ink, accent, CTA and palette, from `src/brand/tokens.json` roles or `BRAND.md`); a colour outside it never reaches the model. The brief describes a complete ad at the master's shape and asks only for its edge margin (see below). `baker canvas set-prompt` refuses these nodes and names `prompt.json`.
 - A **`static_prefilter`** node (tagged `variant_prefilter`) drops variants that fail the deterministic checks before the judge sees them, never all of them; its **judge** (`judge_model`, `judge_prompt`) compares the rest pairwise against the compiled brief when the checks tie (one variant passes through unjudged). A canvas scaffolded with a separate `image_select` judge (tagged `pick`) has it folded into the prefilter on its next run.
-- **Every placement format is made from the chosen picture, and no size is a crop or a paste of another.** The pick goes to exact pixels as the master (an `imagemagick` node tagged `static_master`) at the lead size: the 4:5 when a feed size is delivered, else the square, else the first portrait size, rendered at the model's nearest ratio (3:4 on GPT Image) and filled to Meta's pixels. It is delivered as it is. Every other size (9:16, 1:1, 16:9, …) is **re-laid out** from the master by one request (`image_generate` node `adapt_<size>`, tagged `static_adapt_format`, on the router's `relayout` model, GPT Image 2.5 at quality `max`): the master, the variants' own product and people pictures and the logo file as references and a brief compiled from `prompt.json` for that frame on every run (`static-adapt.ts`): the exact lines and their hierarchy, the same people (same face, same clothes), product and label, logo and none the chosen ad does not show, the words and the button re-scaled for the frame, the frame stated as zones inside its safe area, never the master shrunk onto a background, set as a square or cut. A `size_<size>` node (`static_size`) brings it to Meta's exact pixels (1:1 at 1080×1080; a landscape size the same way). Every 9:16 Reels or Stories covers then passes a free `static_safe_fit` node (`safe_<size>`), which shrinks and shifts the picture just enough (to 85% at most) to bring a word or the logo a few points past the band back inside it; a layout further out passes unchanged and the run lays it out again. The 9:16 is never extended instead (the 4:5 set whole in the clear band, its scene continued to the frame) nor set over its own blurred copy: rated in October 2026, the first showed the ad small inside a continued scene ("text super tiny") and the second "looks bad". The master shape is one decision in the code (`DEFAULT_MASTER_SHAPE`: `relayout`; the `9x16_band` and `4x5_adapt` shapes, kept for the bench, make the sizes by cutting or reframing the master, and the person rating them rejected every such way; a canvas from the retired band master, which extended a square master into the 9:16, is upgraded to the re-layout graph on its next run). Placing the logo file on every size instead of having the model draw it (`logo_composite` nodes) is a scaffold option, off until it is proven on the bench.
+- **Every placement format is made from the chosen picture, and no size is a crop or a paste of another.** The pick goes to exact pixels as the master (an `imagemagick` node tagged `static_master`) at the lead size: the 4:5 when a feed size is delivered, else the square, else the first portrait size, rendered at the model's nearest ratio (3:4 on GPT Image) and filled to Meta's pixels. It is delivered as it is. Every other size (9:16, 1:1, 16:9, …) is **re-laid out** from the master by one request (`image_generate` node `adapt_<size>`, tagged `static_adapt_format`, on the router's `relayout` model, GPT Image 2.5 at quality `xhigh`): the master, the variants' own product and people pictures and the logo file as references and a brief compiled from `prompt.json` for that frame on every run (`static-adapt.ts`): the exact lines and their hierarchy, the same people (same face, same clothes), product and label, logo and none the chosen ad does not show, the words and the button re-scaled for the frame, the frame stated as zones inside its safe area, never the master shrunk onto a background, set as a square or cut. A `size_<size>` node (`static_size`) brings it to Meta's exact pixels (1:1 at 1080×1080; a landscape size the same way). Every 9:16 Reels or Stories covers then passes a free `static_safe_fit` node (`safe_<size>`), which shrinks and shifts the picture just enough (to 85% at most) to bring a word or the logo a few points past the band back inside it; a layout further out passes unchanged and the run lays it out again. The 9:16 is never extended instead (the 4:5 set whole in the clear band, its scene continued to the frame) nor set over its own blurred copy: rated in October 2026, the first showed the ad small inside a continued scene ("text super tiny") and the second "looks bad". The master shape is one decision in the code (`DEFAULT_MASTER_SHAPE`: `relayout`; the `9x16_band` and `4x5_adapt` shapes, kept for the bench, make the sizes by cutting or reframing the master, and the person rating them rejected every such way; a canvas from the retired band master, which extended a square master into the 9:16, is upgraded to the re-layout graph on its next run). Placing the logo file on every size instead of having the model draw it (`logo_composite` nodes) is a scaffold option, off until it is proven on the bench.
 - Readers go by these roles, never by node ids: the run review, `--fix`, `--use-variant`, `--regenerate static_variants`, and the dashboard. A static ad scaffolded before this graph (`ad_a`/`ad_b`/`ad_c`, `pick`, re-lays `ad_4x5`/`ad_9x16`, `fit_4x5`) is **upgraded in place** on its next `validate` or `run`: prompt.json, every asset slot, the model and the formats are kept, the render part becomes this graph, and the creative is not reported as changed by the upgrade. Its old runs stay readable.
 - Detected elements are **never slots to fill**: the source's people, products, animals and badges are its world. The original image stays on the canvas as the record of the source; when `prompt.json` keeps its concept (`concept.verdict` `use` or `adapt`) the first renders are shown it **for its structure only**, and otherwise it never reaches a render. Only the people cast from a company Avatar and the brand's logo become pictures on the canvas; everything else is described in words (a picture of ours is added through a reference's `image`). A **person** with no avatar needs no slot and no casting: the render draws an invented, realistic person from the element's description. With `--avatar <handle>` (for an actor who recurs across ads) the slot ingests the Avatar's identity sheet and `prompt.json` describes that person in the Avatar's own words. A reference's `avatar` names who was cast and draws no face on its own: its face reaches the render only with a picture behind it (the `--avatar` slot, or the identity sheet URL from `baker avatars get <handle>` as the reference's `image`).
 

package/dist/{chunk-GTHLEQHU.js → chunk-MMTPMFFH.js}

@@ -4999,10 +4999,10 @@ var GPT_IMAGE_BRIEF = {
   // (.context/static-review/BENCH.md#E3).
   // quality max: the deterministic checks tie (shippable 9-12% on every step, n=100), but on the
   // E4 gallery (30 rated pairs) the user's win rate is max 69%, xhigh 54%, high 47%, Flare 34%,
-  // and the judge's on the same pairs 73%, 65%, 39%, 31%. The user: results over cost ("use the
-  // best one, not the cheaper"), so the step is the best, at $0.166 per image vs $0.085 for xhigh
-  // and 80 s vs 44 s (BENCH.md#E4, § "E1 v2").
-  params: { n: 1, quality: "max", imageSize: "2K" }
+  // and the judge's on the same pairs 73%, 65%, 39%, 31%. Max costs $0.166 per image vs $0.085
+  // for xhigh and 80 s vs 44 s (BENCH.md#E4, § "E1 v2"); a round renders twice in a row (the
+  // options, then each size), so three ads took most of an hour. Oct 2026 the user chose xhigh.
+  params: { n: 1, quality: "xhigh", imageSize: "2K" }
 };
 var GEMINI_IMAGE_BRIEF = {
   // "Use Natural Language & Full Sentences… as if you were briefing a human artist", and
@@ -8674,7 +8674,7 @@ function sizeNode(s, adaptFormats) {
   const index = adaptFormats.indexOf(s.adaptFormat ?? s.format);
   return magick(id, STATIC_SIZE_ROLE, `$ref:adapt.images#${index}`, fillArgs(s.size));
 }
-var RELAYOUT_QUALITY = "max";
+var RELAYOUT_QUALITY = "xhigh";
 function relayoutNodes(g) {
   return g.plan.sizes.filter((s) => s.from === "relayout").map((s) => relayoutNode(s, g));
 }
@@ -18564,4 +18564,4 @@ export {
   defaultRegistry,
   createEngineFromEnv
 };
-//# sourceMappingURL=chunk-GTHLEQHU.js.map
+//# sourceMappingURL=chunk-MMTPMFFH.js.map