@shop-os/modelkit 0.2.20 → 0.2.21

package/CHANGELOG.md

@@ -9,6 +9,53 @@ header in the release PR.
 
 ## [Unreleased]
 
+## [0.2.21] - Unreleased
+
+### Added
+- `capabilities.lastFrameRequiresFirstFrame` (boolean): a video row that takes an end frame only together with a start frame. On such a row, `resolveVideoOptions` in `error` mode refuses an end frame without a start frame with a non-retryable `missing_input`; in `coerce` mode the lone end frame becomes the start frame, with an `option_ignored` warning ("end frame used as start frame: <model> needs a start frame with an end frame"). Set on `veo-3.1` ("must be used in combination with the image parameter") and `gemini-omni-1.1-flash` (`<LAST_FRAME>` "must be used with `<FIRST_FRAME>`"). Not set on `kling-v3-omni-video` (`end_image` "Requires start_image"): error-mode callers get `missing_input`, and SpaceLab's vibe-triggers resolves replicate and google-genai rows in error mode, so the flag would refuse its end-frame-only Kling Omni nodes. Without the flag the Replicate adapter keeps sending a lone end frame as `start_image`.
+- `gemini-omni-1.1-flash` takes an end frame (`supports: ["videoExtension", "lastFrame"]`). The adapter sends the start image, then the end image, then any references, and prefixes the prompt with the docs' role tags `<FIRST_FRAME> <LAST_FRAME>`. A chained extension drops the end frame with a warning, as it drops the other conditioning media.
+- Higgsfield `kling-3.0` sound switch: `generateAudio` maps to the endpoint's `sound: "on" | "off"`. The row is now `audio: "optional"`, `supports` `generateAudio`, default `generateAudio: true` (the provider default is on).
+- `kling-v3-omni-video` offers `4k`: the Replicate adapter maps `resolution: "4k"` to `mode: "4k"` (4K does not take a reference video, which the adapter never sends).
+- `constraints.prompt.maxChars`: 2500 on `kling-2.5`, `kling-3.0` and `kling-v3-omni-video` (Higgsfield 3.0 truncates past 2,500; Replicate's Omni says "Max 2500 characters"; Higgsfield 2.5 documents none, the same model on fal and Kling's API caps it at 2,500), 3000 on `veo-3.1` (1,024 text tokens at about 2.9 characters per token; a CJK prompt can hit the token limit sooner). Each keeps its `numbered` reference syntax.
+
+### Changed
+- Higgsfield Kling adapter sends each endpoint only the fields its schema has. `kling-video/v2.5-turbo/pro/image-to-video` gets no `aspect_ratio` and no `last_image_url`; `kling-video/v3.0/pro/image-to-video` gets no `aspect_ratio` and no `negative_prompt`. A dropped caller value becomes an `option_ignored` warning. An unlisted Higgsfield Kling endpoint keeps the fields it was sent before.
+- `kling-2.5`: no `aspectRatios`, no ratio default, `supports` loses `lastFrame`.
+- `kling-3.0`: no `aspectRatios`, no ratio default, `supports` loses `negativePrompt` (and the default negative prompt).
+- `kling-v3-omni-video`: `aspectRatios` loses `adaptive` (not in Replicate's enum; "Auto" meant the provider's 16:9); `resolutions` gains `4k`; default `generateAudio: true`.
+- `seedance-2.0`, `seedance-2.0-fast`, `seedance-2.5`: `supports` loses `seed`. BytePlus accepts and echoes `seed` on 2.x, but two identical seeded tasks gave different videos (SSIM 0.83, 2026-10-01). The ark adapter still sends `seed` for rows that support it (Seedream and others).
+- `seedance-2.5`: default `generateAudio: true` (sound on by default where it is optional).
+- `minimax-h3-max`, `minimax-h3-max-reference`: `resolutions` gain `1080p` (sent as fal's `1080P`).
+- `minimax-h3`, `minimax-h3-max`, `minimax-h3-reference`, `minimax-h3-max-reference`: `ratioIgnoredWith: ["firstFrame"]`. A start frame routes to fal's image-to-video endpoint, which has no `aspect_ratio`.
+- `grok-imagine`: `references.images` 7 (was 5); the xAI adapter caps at 7 ("A maximum of 7 reference images can be provided per request"). The description still says 5 because it is pinned to the Cowork catalog fixture.
+
+### Deferred
+- `veo-3.1` stays at `durations: [8]`. Veo also takes 4 s and 6 s, but must be 8 s at 1080p or 4k and with reference images, and the kit has no rule that forces a duration from other options. Offering `[4, 6, 8]` now would snap SpaceLab's stored Veo nodes (default 5 s) to 4 s, which Veo rejects at 1080p/4k and which halves 720p clips. Ship it with a kit rule that forces 8 s in those cases (`veo-3.1-fast` and `veo-3.1-lite` already list 4/6/8 and have this gap).
+
+### Live rows to edit after publish
+Apps read the pro-backend `model_catalog` rows, not this file. Before the live edit, add `credit_pricing_video` rows for `kling-v3-omni-video` at `4k` and for `minimax-h3-max` / `minimax-h3-max-reference` at `1080p`: without them those options fail to quote. At the version bump in shopos-ui, update vibe-api's pinned pricing producer contract (`contracts/backend-pricing-contract.json`, checked by `producer-contract.test.ts` against the bundled rows), which gains the same `kling-v3-omni-video`/`4k` and `minimax-h3-max`/`1080p` keys.
+
+Preconditions:
+- Remove the `kling-2.5` and `kling-3.0` aspect ratios only after cowork with Shop-OS/shopos-backend-ts#1197 (including its ratio-less projection fix) is deployed. Before that, #1197's `?frames=1` projection 500s on a ratio-less row.
+- Set `capabilities.lastFrameRequiresFirstFrame` through pro-backend's admin catalog API. Don't save the `veo-3.1` or `gemini-omni-1.1-flash` rows in pro-admin until its row form keeps unknown capability keys and has the checkbox (fix in Shop-OS/shopos-ui#2711); today it drops the whole `capabilities` object.
+
+Behaviour after the edit:
+- `seedance-2.5` and `kling-v3-omni-video` default to sound on, and chat and Creative Director send the row default, so their videos gain sound too. Credits are unchanged until audio pricing lands. Kling V3 Omni with audio costs $0.224/s at standard (720p) and $0.28/s at pro (1080p) on Replicate, against $0.168/s and $0.224/s without: check the `credit_pricing_video` margin.
+- A SpaceLab Kling 3.0 node with a stored `generate_audio: false` goes silent after the `kling-3.0` edit, where it had sound before.
+
+After pro-backend is on 0.2.21, in each environment:
+- `kling-2.5`: remove `constraints.aspectRatios`; remove `defaults.aspectRatio`; `constraints.supports` → `["cfgScale", "negativePrompt"]`; `constraints.prompt` → `{ "referenceSyntax": "numbered", "maxChars": 2500 }`; `display.description` → the new text (no end image).
+- `kling-3.0`: remove `constraints.aspectRatios`; remove `defaults.aspectRatio` and `defaults.negativePrompt`; `constraints.supports` → `["cfgScale", "lastFrame", "generateAudio"]`; `constraints.audio` → `"optional"`; `defaults.generateAudio` → `true`; `constraints.prompt` → `{ "referenceSyntax": "numbered", "maxChars": 2500 }`. Higgsfield lists this endpoint at $0.112-$0.168/s without labelling the sound split; check whether `credit_pricing_video` should price sound on and off separately.
+- `kling-v3-omni-video`: `constraints.aspectRatios` → `["16:9", "9:16", "1:1"]`; `constraints.resolutions` → `["720p", "1080p", "4k"]`; `defaults.generateAudio` → `true`; `constraints.prompt` → `{ "referenceSyntax": "numbered", "maxChars": 2500 }`; `display.description` → "... up to 4K". Needs a `credit_pricing_video` row for `kling-v3-omni-video` at `4k` first (Replicate: $0.42/s, with or without audio).
+- `veo-3.1`: `constraints.prompt` → `{ "referenceSyntax": "numbered", "maxChars": 3000 }`; `capabilities.lastFrameRequiresFirstFrame` → `true`.
+- `gemini-omni-1.1-flash`: `constraints.supports` → `["videoExtension", "lastFrame"]`; `capabilities.lastFrameRequiresFirstFrame` → `true`.
+- `seedance-2.0`, `seedance-2.0-fast`: `constraints.supports` → `["generateAudio", "lastFrame", "watermark"]`.
+- `seedance-2.5`: `constraints.supports` → `["generateAudio", "lastFrame", "watermark", "taskType"]`; `defaults.generateAudio` → `true`.
+- `minimax-h3`, `minimax-h3-reference`: `constraints.ratioIgnoredWith` → `["firstFrame"]`.
+- `minimax-h3-max`, `minimax-h3-max-reference`: `constraints.resolutions` → `["480p", "768p", "1080p"]`; `constraints.ratioIgnoredWith` → `["firstFrame"]`; `display.description` of `minimax-h3-max` → "... up to 1080p". Needs `credit_pricing_video` rows for both at `1080p`.
+- `grok-imagine`: `constraints.references.images` → `7`.
+- `p-video-2-pro` (pro-admin-only, not in this file): `constraints.aspectRatios` without `adaptive` (Replicate's enum has none); `constraints.audio` → `"always"` (still outstanding from 0.2.20).
+
 ## [0.2.20] - Unreleased
 
 ### Added

package/dist/catalog.cjs

@@ -1306,12 +1306,14 @@ var TEXT_MODELS = [
 var VIDEO_USAGE = { unit: "seconds" };
 var SEEDANCE_RATIOS = ["21:9", "16:9", "4:3", "1:1", "3:4", "9:16", "adaptive"];
 var KLING_RATIOS = ["16:9", "9:16", "1:1", "adaptive"];
+var KLING_OMNI_RATIOS = ["16:9", "9:16", "1:1"];
 var OMNI_RATIOS = ["16:9", "9:16"];
 var GROK_RATIOS = ["1:1", "16:9", "9:16", "4:3", "3:4", "3:2", "2:3"];
 var H3_RATIOS = ["21:9", "16:9", "4:3", "1:1", "3:4", "9:16"];
 var H3_REFERENCE_RATIOS = ["adaptive", ...H3_RATIOS];
 var NEGATIVE_PROMPT = "blur, distort, and low quality";
 var FRAMES_EXCLUDE_REFERENCES = { framesExcludeReferences: true };
+var KLING_PROMPT = { referenceSyntax: "numbered", maxChars: 2500 };
 var BYTEDANCE_LOGO = "https://d31vvfpk2ppdyq.cloudfront.net/assets/image/3bba5d81-4e57-4e71-911b-cdfa05acca2c_compressed.png";
 var GOOGLE_LOGO2 = "https://d31vvfpk2ppdyq.cloudfront.net/assets/image/192485d2-a76a-4657-9c06-930b2c70badd_compressed.png";
 var VIDEO_MODELS = [
@@ -1332,9 +1334,13 @@ var VIDEO_MODELS = [
     constraints: {
       aspectRatios: [...OMNI_RATIOS],
       resolutions: ["720p", "1080p", "4k"],
+      // Veo also takes 4 and 6, but must be 8 at 1080p/4k or with reference images;
+      // [8] until the kit has a rule that forces it there (CHANGELOG 0.2.21).
       durations: [8],
       references: { images: 3 },
       supports: ["lastFrame", "seed"],
+      // 1,024 text tokens; 3,000 characters assumes ~2.9 characters per token.
+      prompt: { referenceSyntax: "numbered", maxChars: 3e3 },
       audio: "always"
     },
     defaults: { durationSeconds: 8, resolution: "720p", aspectRatio: "16:9" },
@@ -1345,7 +1351,8 @@ var VIDEO_MODELS = [
       description: "Google's highest-quality video generation model. Supports up to 4K resolution and 8-second clips with reference image conditioning.",
       logo: GOOGLE_LOGO2
     },
-    capabilities: { ...FRAMES_EXCLUDE_REFERENCES }
+    // Veo's lastFrame "must be used in combination with the image parameter".
+    capabilities: { ...FRAMES_EXCLUDE_REFERENCES, lastFrameRequiresFirstFrame: true }
   },
   {
     id: "veo-3.1-fast",
@@ -1465,7 +1472,8 @@ var VIDEO_MODELS = [
       // extension can upload a <= 10 s source with task "extend"; later ones
       // chain by `previousInteractionId` (the API rejects the id together
       // with a video task, so those calls carry the prompt alone).
-      supports: ["videoExtension"],
+      // First-to-last-frame interpolation: `<LAST_FRAME>` goes with `<FIRST_FRAME>`.
+      supports: ["videoExtension", "lastFrame"],
       extension: { segmentSeconds: 10, maxTotalSeconds: 40 },
       // generateOmniVideo sends MediaRef.label as a text part beside each image.
       prompt: { referenceSyntax: "caption" },
@@ -1479,7 +1487,7 @@ var VIDEO_MODELS = [
       logo: GOOGLE_LOGO2,
       description: "Google Gemini Omni Flash 1.1 conversational video generation, editing and extension; text, image, and up to 10-second video input, 3-10 second output at up to 4K and 24 FPS, extendable to 40 seconds."
     },
-    capabilities: { acceptsSourceVideo: true }
+    capabilities: { acceptsSourceVideo: true, lastFrameRequiresFirstFrame: true }
   },
   {
     // vibe-triggers and Space Builder drop the ratio when a frame is wired; Cowork always sends it.
@@ -1505,7 +1513,8 @@ var VIDEO_MODELS = [
       },
       sourceVideoSeconds: { min: 2, max: 15, inReferenceTotal: true },
       ratioIgnoredWith: ["firstFrame", "sourceVideo"],
-      supports: ["generateAudio", "seed", "lastFrame", "watermark"],
+      // No seed: 2.x echoes it but does not honour it (two seeded runs differed, 2026-10-01).
+      supports: ["generateAudio", "lastFrame", "watermark"],
       audio: "optional"
     },
     defaults: {
@@ -1545,7 +1554,8 @@ var VIDEO_MODELS = [
       },
       sourceVideoSeconds: { min: 2, max: 15, inReferenceTotal: true },
       ratioIgnoredWith: ["firstFrame", "sourceVideo"],
-      supports: ["generateAudio", "seed", "lastFrame", "watermark"],
+      // No seed: 2.x echoes it but does not honour it (two seeded runs differed, 2026-10-01).
+      supports: ["generateAudio", "lastFrame", "watermark"],
       audio: "optional"
     },
     defaults: {
@@ -1564,7 +1574,6 @@ var VIDEO_MODELS = [
     capabilities: { acceptsSourceVideo: true, ...FRAMES_EXCLUDE_REFERENCES }
   },
   {
-    // generateAudio defaults to false here (trigger-workflows); Space Builder defaults it to true.
     // omni_reference_task_type edit/extend: adaptive ratio, edit sends duration -1.
     id: "seedance-2.5",
     aliases: ["seedance-2-5", "dreamina-seedance-2-5-260628"],
@@ -1586,14 +1595,15 @@ var VIDEO_MODELS = [
       },
       sourceVideoSeconds: { min: 2, max: 30, editMin: 4, inReferenceTotal: true },
       ratioIgnoredWith: ["firstFrame", "sourceVideo"],
-      supports: ["generateAudio", "seed", "lastFrame", "watermark", "taskType"],
+      // No seed: 2.x echoes it but does not honour it (two seeded runs differed, 2026-10-01).
+      supports: ["generateAudio", "lastFrame", "watermark", "taskType"],
       audio: "optional"
     },
     defaults: {
       durationSeconds: 5,
       resolution: "720p",
       aspectRatio: "16:9",
-      generateAudio: false
+      generateAudio: true
     },
     usage: { ...VIDEO_USAGE },
     pricingKind: "video_per_second",
@@ -1605,7 +1615,6 @@ var VIDEO_MODELS = [
     capabilities: { acceptsSourceVideo: true, ...FRAMES_EXCLUDE_REFERENCES }
   },
   {
-    // Cowork allows 7 reference images and durations 4/6/8/10; Space Builder allows 5 and 1-15s.
     id: "grok-imagine",
     aliases: ["grok-imagine-video", "grokImagine"],
     label: "Grok Imagine",
@@ -1619,7 +1628,8 @@ var VIDEO_MODELS = [
       resolutions: ["480p", "720p"],
       durations: { min: 1, max: 15 },
       durationPresets: [4, 6, 8, 10],
-      references: { images: 5 },
+      // xAI: "A maximum of 7 reference images can be provided per request."
+      references: { images: 7 },
       supports: [],
       // xAI can switch sound off (generate_audio); xai-video.ts never sends it.
       audio: "always"
@@ -1644,16 +1654,16 @@ var VIDEO_MODELS = [
     providerModelId: "kling-video/v2.5-turbo/pro/image-to-video",
     lifecycle: "active",
     exposure: ["cowork", "space_builder"],
+    // The endpoint has no aspect_ratio or last_image_url: the start image sets the framing.
     constraints: {
-      aspectRatios: [...KLING_RATIOS],
       durations: [5, 10],
       requires: ["firstFrame"],
-      supports: ["cfgScale", "negativePrompt", "lastFrame"],
+      supports: ["cfgScale", "negativePrompt"],
+      prompt: KLING_PROMPT,
       audio: "never"
     },
     defaults: {
       durationSeconds: 5,
-      aspectRatio: "16:9",
       cfgScale: 0.5,
       negativePrompt: NEGATIVE_PROMPT
     },
@@ -1662,7 +1672,7 @@ var VIDEO_MODELS = [
     display: {
       provider: "kuaishou",
       logo: "https://cdn.jsdelivr.net/npm/@lobehub/icons-static-png@1.97.0/light/kling-color.png",
-      description: "Kuaishou Kling 2.5 Turbo Pro image-to-video; animates a single input image from a text prompt with an optional end image, 5- or 10-second clips."
+      description: "Kuaishou Kling 2.5 Turbo Pro image-to-video; animates a single input image from a text prompt, 5- or 10-second clips."
     },
     capabilities: { ...FRAMES_EXCLUDE_REFERENCES }
   },
@@ -1676,19 +1686,18 @@ var VIDEO_MODELS = [
     providerModelId: "kling-video/v3.0/pro/image-to-video",
     lifecycle: "active",
     exposure: ["cowork", "space_builder"],
+    // The endpoint has no aspect_ratio or negative_prompt; sound is "on"/"off" (default on).
     constraints: {
-      aspectRatios: [...KLING_RATIOS],
       durations: { min: 3, max: 15 },
       requires: ["firstFrame"],
-      supports: ["cfgScale", "negativePrompt", "lastFrame"],
-      // Higgsfield has a `sound` switch (default on); KlingCreate doesn't send it.
-      audio: "always"
+      supports: ["cfgScale", "lastFrame", "generateAudio"],
+      prompt: KLING_PROMPT,
+      audio: "optional"
     },
     defaults: {
       durationSeconds: 5,
-      aspectRatio: "16:9",
       cfgScale: 0.5,
-      negativePrompt: NEGATIVE_PROMPT
+      generateAudio: true
     },
     usage: { ...VIDEO_USAGE },
     pricingKind: "video_per_second",
@@ -1700,7 +1709,7 @@ var VIDEO_MODELS = [
     capabilities: { ...FRAMES_EXCLUDE_REFERENCES }
   },
   {
-    // mode standard/pro selects 720p/1080p; the endpoint takes no separate resolution.
+    // mode standard/pro/4k selects 720p/1080p/4K; the endpoint takes no separate resolution.
     // Replicate: "Max 7 [images] without video, 4 with video"; 4K does not support
     // reference_video. No `videos` limit yet: the adapter sends no reference video,
     // so `imagesWithVideo` stays dormant until one is wired through.
@@ -1713,22 +1722,25 @@ var VIDEO_MODELS = [
     lifecycle: "active",
     exposure: ["cowork", "space_builder"],
     constraints: {
-      aspectRatios: [...KLING_RATIOS],
-      resolutions: ["720p", "1080p"],
+      aspectRatios: [...KLING_OMNI_RATIOS],
+      resolutions: ["720p", "1080p", "4k"],
       durations: { min: 3, max: 15 },
       references: { images: 7, imagesWithVideo: 4 },
       supports: ["generateAudio", "lastFrame"],
+      prompt: KLING_PROMPT,
       audio: "optional"
     },
-    defaults: { durationSeconds: 5, aspectRatio: "16:9", generateAudio: false, mode: "pro" },
+    defaults: { durationSeconds: 5, aspectRatio: "16:9", generateAudio: true, mode: "pro" },
     usage: { unit: "compute_seconds" },
     pricingKind: "video_per_second",
     display: {
       provider: "kuaishou",
       logo: "https://cdn.jsdelivr.net/npm/@lobehub/icons-static-png@1.97.0/light/kling-color.png",
-      description: "Kuaishou Kling Video 3.0 Omni unified generation and editing from text, start/end images, up to 7 reference images, and reference video; up to 15 seconds at up to 1080p."
-    },
-    capabilities: {}
+      description: "Kuaishou Kling Video 3.0 Omni unified generation and editing from text, start/end images, up to 7 reference images, and reference video; up to 15 seconds at up to 4K."
+    }
+    // end_image "Requires start_image", but no lastFrameRequiresFirstFrame: SpaceLab resolves
+    // replicate rows in error mode, so the flag would refuse its end-frame-only nodes, which
+    // the adapter sends as start_image.
   },
   {
     id: "minimax-h3",
@@ -1743,6 +1755,8 @@ var VIDEO_MODELS = [
       aspectRatios: [...H3_RATIOS],
       resolutions: ["480p", "768p", "2k", "4k"],
       durations: { min: 5, max: 15 },
+      // fal's image-to-video endpoint has no aspect_ratio; the frame sets it.
+      ratioIgnoredWith: ["firstFrame"],
       supports: ["seed", "promptExpansion", "lastFrame", "safetyChecker"],
       audio: "always"
     },
@@ -1772,8 +1786,10 @@ var VIDEO_MODELS = [
     exposure: ["cowork", "space_builder"],
     constraints: {
       aspectRatios: [...H3_RATIOS],
-      resolutions: ["480p", "768p"],
+      resolutions: ["480p", "768p", "1080p"],
       durations: { min: 5, max: 15 },
+      // fal's image-to-video endpoint has no aspect_ratio; the frame sets it.
+      ratioIgnoredWith: ["firstFrame"],
       supports: ["seed", "promptExpansion", "lastFrame", "safetyChecker"],
       audio: "always"
     },
@@ -1788,7 +1804,7 @@ var VIDEO_MODELS = [
     display: {
       provider: "minimax",
       logo: "https://d31vvfpk2ppdyq.cloudfront.net/assets/image/f4c2e906-28b7-4a4a-b491-df8fef7808ed.png",
-      description: "High-speed MiniMax H3 Max variant for text-to-video and image-to-video (first/last frame) only, up to 768p."
+      description: "High-speed MiniMax H3 Max variant for text-to-video and image-to-video (first/last frame) only, up to 1080p."
     },
     capabilities: { ...FRAMES_EXCLUDE_REFERENCES }
   },
@@ -1805,6 +1821,8 @@ var VIDEO_MODELS = [
       aspectRatios: [...H3_REFERENCE_RATIOS],
       resolutions: ["480p", "768p", "2k", "4k"],
       durations: { min: 5, max: 15 },
+      // A start frame routes to image-to-video, which has no aspect_ratio.
+      ratioIgnoredWith: ["firstFrame"],
       references: {
         images: 9,
         videos: 3,
@@ -1844,8 +1862,10 @@ var VIDEO_MODELS = [
     exposure: ["cowork", "space_builder"],
     constraints: {
       aspectRatios: [...H3_REFERENCE_RATIOS],
-      resolutions: ["480p", "768p"],
+      resolutions: ["480p", "768p", "1080p"],
       durations: { min: 5, max: 15 },
+      // A start frame routes to image-to-video, which has no aspect_ratio.
+      ratioIgnoredWith: ["firstFrame"],
       references: {
         images: 9,
         videos: 3,
@@ -2403,6 +2423,7 @@ var CatalogDisplaySchema = import_zod3.z.looseObject({
 var CatalogCapabilitiesSchema = import_zod3.z.looseObject({
   acceptsSourceVideo: import_zod3.z.boolean().optional(),
   framesExcludeReferences: import_zod3.z.boolean().optional(),
+  lastFrameRequiresFirstFrame: import_zod3.z.boolean().optional(),
   reasoning: import_zod3.z.boolean().optional(),
   vision: import_zod3.z.boolean().optional(),
   contextWindow: import_zod3.z.number().optional()
@@ -2984,6 +3005,24 @@ function resolveVideoOptions(model, request, onUnsupported = "error") {
     });
     delete merged.sourceVideo;
   }
+  if (merged.lastFrame !== void 0 && merged.firstFrame === void 0 && model.capabilities?.lastFrameRequiresFirstFrame) {
+    if (onUnsupported === "error") {
+      throw new ModelKitError({
+        code: "missing_input",
+        message: `${model.id} takes an end frame only together with a start frame`,
+        model: model.id,
+        retryable: false
+      });
+    }
+    warnings.push({
+      code: "option_ignored",
+      field: "lastFrame",
+      to: "firstFrame",
+      message: `end frame used as start frame: ${model.id} needs a start frame with an end frame`
+    });
+    merged.firstFrame = merged.lastFrame;
+    delete merged.lastFrame;
+  }
   const hasFrames = merged.firstFrame !== void 0 || merged.lastFrame !== void 0;
   const hasRefs = compact(merged.referenceImages).length + compact(merged.referenceVideos).length + compact(merged.referenceAudios).length > 0;
   if (hasFrames && hasRefs && model.capabilities?.framesExcludeReferences) {