@shop-os/modelkit 0.7.7 → 0.7.10

package/dist/index.js

@@ -5129,8 +5129,10 @@ var RAY_RESOLUTIONS = ["360p", "720p", "1080p"];
 var RAY_PROMPT = { referenceSyntax: "numbered", maxChars: 6e3 };
 var LUMA_VIDEO_MODELS = [
   {
-    // https://docs.agents.lumalabs.ai: type "video" with a fixed 5 s duration;
-    // firstFrame and lastFrame ride as start_frame and end_frame.
+    // https://docs.agents.lumalabs.ai: type "video" with a 5 s or 10 s duration;
+    // firstFrame and lastFrame ride as start_frame and end_frame. Luma bills a
+    // 10 s clip at 3x the 5 s price; pro-backend prices it that way, so this row
+    // must not offer 10 s before that pricing is deployed (ENG-3324).
     id: "ray-3.2",
     aliases: [],
     label: "Ray 3.2",
@@ -5142,7 +5144,7 @@ var LUMA_VIDEO_MODELS = [
     constraints: {
       aspectRatios: ["adaptive", ...RAY_RATIOS],
       resolutions: [...RAY_RESOLUTIONS],
-      durations: [5],
+      durations: [5, 10],
       supports: ["lastFrame"],
       prompt: { ...RAY_PROMPT },
       audio: "never"
@@ -5152,7 +5154,7 @@ var LUMA_VIDEO_MODELS = [
     pricingKind: "video_per_second",
     display: {
       provider: "luma",
-      description: "Luma Ray 3.2 text- and image-to-video with optional first and last frames, 5-second clips at up to 1080p.",
+      description: "Luma Ray 3.2 text- and image-to-video with optional first and last frames, 5- or 10-second clips at up to 1080p.",
       logo: LUMA_LOGO2
     },
     capabilities: {}
@@ -5662,38 +5664,38 @@ var VIDEO_MODELS = [
     capabilities: { ...FRAMES_EXCLUDE_REFERENCES }
   },
   {
-    // 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.
+    // fal Kling O3 (adapters/fal-kling-omni.ts): mode standard/pro/4k selects 720p/1080p/4K
+    // in the endpoint path. Moved off Replicate kwaivgi/kling-v3-omni-video, which fal sells
+    // at half the price; the id and inputs stay. The Replicate mapper stays for deployed
+    // model_catalog rows not yet switched. Frames and reference images share fal's cap of 7.
+    // Reference video, multi-shot and keep-sound were never wired, so stay off.
     id: "kling-v3-omni-video",
     aliases: ["kwaivgi/kling-v3-omni-video", "kling-v3-omni"],
     label: "Kling V3 Omni",
     type: "video",
-    adapter: "replicate",
-    providerModelId: "kwaivgi/kling-v3-omni-video",
+    adapter: "fal",
+    providerModelId: "fal-ai/kling-video/o3",
     lifecycle: "active",
     exposure: ["cowork", "space_builder"],
     constraints: {
       aspectRatios: [...KLING_OMNI_RATIOS],
       resolutions: ["720p", "1080p", "4k"],
       durations: { min: 3, max: 15 },
-      references: { images: 7, imagesWithVideo: 4 },
+      references: { images: 7, videos: 0, audios: 0 },
       supports: ["generateAudio", "lastFrame"],
       prompt: KLING_PROMPT2,
       audio: "optional"
     },
     defaults: { durationSeconds: 5, aspectRatio: "16:9", generateAudio: true, mode: "pro" },
-    usage: { unit: "compute_seconds" },
+    usage: { ...VIDEO_USAGE3 },
     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 4K."
+      description: "Kuaishou Kling Video 3.0 Omni generation from text, start/end images and up to 7 reference images, with optional native audio; 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.
+    // No lastFrameRequiresFirstFrame: fal image-to-video needs a start frame, so the adapter
+    // sends a lone end frame as the start frame, as the Replicate model took it.
   },
   {
     id: "minimax-h3",
@@ -9308,11 +9310,11 @@ function toolInput(ctx, req, output) {
 }
 var EXR_DEFAULT_PROMPT = "The same video in high dynamic range, with unchanged content, motion, framing and timing.";
 async function submitFalVideoTool(req, ctx, opts) {
-  const plan3 = planVideoTool(ctx.model, req);
-  if (!plan3.ok) {
+  const plan4 = planVideoTool(ctx.model, req);
+  if (!plan4.ok) {
     throw new ModelKitError({
       code: "unsupported_option",
-      message: plan3.error,
+      message: plan4.error,
       provider: "fal",
       model: ctx.model.id,
       retryable: false
@@ -9320,7 +9322,7 @@ async function submitFalVideoTool(req, ctx, opts) {
   }
   const target = new URL(`${FAL_QUEUE_ORIGIN}/${ctx.model.providerModelId}`);
   if (opts.webhookUrl !== void 0) target.searchParams.set("fal_webhook", opts.webhookUrl);
-  const res = await queueFetch(ctx, target.toString(), "POST", toolInput(ctx, req, plan3.output));
+  const res = await queueFetch(ctx, target.toString(), "POST", toolInput(ctx, req, plan4.output));
   const parsed = submitResponseSchema.safeParse(await readJson(ctx, res));
   if (!parsed.success) {
     throw new ModelKitError({
@@ -9338,7 +9340,7 @@ async function submitFalVideoTool(req, ctx, opts) {
     statusUrl: parsed.data.status_url,
     responseUrl: parsed.data.response_url,
     cancelUrl: parsed.data.cancel_url,
-    output: plan3.output
+    output: plan4.output
   };
 }
 async function falVideoToolStatus(sub, ctx) {
@@ -10027,14 +10029,14 @@ function planFalImage(req, providerModelId, model) {
   const exact = falExactSize(spec, req);
   const call = { label: spec.label, req, refs, edit: edit2, ...exact !== void 0 && { exact } };
   const built = family.build(call, model);
-  const plan3 = {
+  const plan4 = {
     endpoint: edit2 && family.editEndpoint !== void 0 ? family.editEndpoint : id,
     input: built.input,
     output: family.output ?? "images"
   };
-  if (built.imageQuality !== void 0) plan3.imageQuality = built.imageQuality;
-  if (built.imageSize !== void 0) plan3.imageSize = built.imageSize;
-  return plan3;
+  if (built.imageQuality !== void 0) plan4.imageQuality = built.imageQuality;
+  if (built.imageSize !== void 0) plan4.imageSize = built.imageSize;
+  return plan4;
 }
 var FalFileSchema = z14.object({ url: z14.string(), content_type: z14.string().nullish() });
 var ImagesPayloadSchema = z14.object({ images: z14.array(FalFileSchema) });
@@ -10056,18 +10058,18 @@ async function downloadFalFile(ctx, file) {
   const mime = res.headers.get("content-type") ?? file.content_type ?? "";
   return { bytes: new Uint8Array(await res.arrayBuffer()), mime };
 }
-function filesOf(payload, plan3) {
-  if (plan3.output === "image") {
+function filesOf(payload, plan4) {
+  if (plan4.output === "image") {
     const parsed2 = ImagePayloadSchema.safeParse(payload);
     return parsed2.success ? [parsed2.data.image] : [];
   }
   const parsed = ImagesPayloadSchema.safeParse(payload);
   return parsed.success ? parsed.data.images : [];
 }
-async function collectFalPlanImages(payload, plan3, ctx, req) {
-  const files = filesOf(payload, plan3);
+async function collectFalPlanImages(payload, plan4, ctx, req) {
+  const files = filesOf(payload, plan4);
   if (files.length === 0) {
-    throw fail2(ctx.model.id, `fal ${plan3.endpoint} returned no image`, "provider_rejected");
+    throw fail2(ctx.model.id, `fal ${plan4.endpoint} returned no image`, "provider_rejected");
   }
   const images = await Promise.all(
     files.map(async (file) => {
@@ -10082,18 +10084,18 @@ async function collectFalPlanImages(payload, plan3, ctx, req) {
     unit: "images",
     quantity: images.length
   };
-  const billedSize = plan3.imageSize ?? req.size;
+  const billedSize = plan4.imageSize ?? req.size;
   if (billedSize !== void 0 && ctx.model.type === "image") usage3.imageSize = billedSize;
-  if (plan3.imageQuality !== void 0) usage3.imageQuality = plan3.imageQuality;
+  if (plan4.imageQuality !== void 0) usage3.imageQuality = plan4.imageQuality;
   const result = { images, usage: usage3, raw: payload };
   const remote = files.find((f) => !f.url.startsWith("data:"));
   if (remote !== void 0) result.outputUrl = remote.url;
   return result;
 }
 async function generateFalFamilyImage(req, ctx, run) {
-  const plan3 = planFalImage(req, ctx.model.providerModelId, ctx.model.id);
-  const payload = await run(plan3.endpoint, plan3.input);
-  return collectFalPlanImages(payload, plan3, ctx, req);
+  const plan4 = planFalImage(req, ctx.model.providerModelId, ctx.model.id);
+  const payload = await run(plan4.endpoint, plan4.input);
+  return collectFalPlanImages(payload, plan4, ctx, req);
 }
 
 // src/adapters/fal-mask.ts
@@ -11191,10 +11193,50 @@ var KLING_TIERS = [
   { mode: "pro", resolution: "1080p" },
   { mode: "4k", resolution: "4k" }
 ];
+var KLING_O3_OMNI = "fal-ai/kling-video/o3";
 function klingTier(req) {
   return KLING_TIERS.find((tier) => tier.resolution === req.resolution) ?? KLING_TIERS.find((tier) => tier.mode === req.mode) ?? KLING_TIERS[0];
 }
 
+// src/adapters/fal-kling-omni.ts
+function plan(req, ctx) {
+  const tier = klingTier(req);
+  const duration = wholeSeconds(req, ctx);
+  const input = {
+    prompt: req.prompt.trim(),
+    duration: String(duration),
+    generate_audio: req.generateAudio ?? true
+  };
+  const start = req.firstFrame ?? req.lastFrame;
+  const end = req.firstFrame ? req.lastFrame : void 0;
+  const images = urls(req.referenceImages);
+  let task;
+  if (images.length > 0) {
+    task = "reference-to-video";
+    input["image_urls"] = images;
+    if (start) input["start_image_url"] = refToUrl(start);
+    if (end) input["end_image_url"] = refToUrl(end);
+  } else if (start) {
+    task = "image-to-video";
+    input["image_url"] = refToUrl(start);
+    if (end) input["end_image_url"] = refToUrl(end);
+  } else {
+    task = "text-to-video";
+  }
+  const ratio = concreteRatio(req);
+  if (ratio !== void 0 && start === void 0) input["aspect_ratio"] = ratio;
+  return {
+    endpoint: `${KLING_O3_OMNI}/${tier.mode}/${task}`,
+    input,
+    resolution: tier.resolution,
+    seconds: requestedSeconds(duration)
+  };
+}
+var klingOmniFamily = {
+  ids: /* @__PURE__ */ new Set([KLING_O3_OMNI]),
+  plan
+};
+
 // src/resolve-video-rules.ts
 var MOTION_PRO2 = "fal-ai/kling-video/v3/pro/motion-control";
 var MOTION_STANDARD2 = "fal-ai/kling-video/v3/standard/motion-control";
@@ -11339,6 +11381,7 @@ var RULES = /* @__PURE__ */ new Map([
   [MOTION_PRO2, motionControl2],
   [MOTION_STANDARD2, motionControl2],
   [O3_REFERENCE2, o3Reference2],
+  [KLING_O3_OMNI, o3Reference2],
   [O3_EDIT2, (req) => needsSource(req) ?? needsPrompt(req)],
   [H3_TURBO, h3Turbo],
   [MINIMAX_H3_INSERT, h3Insert],
@@ -11368,13 +11411,14 @@ function h3EndFrameRatioWarning(model, ratio) {
     message: `${model.id} follows the end frame's aspect ratio`
   };
 }
-function isReplicateKling(model) {
+function takesKlingMode(model) {
+  if (model.adapter === "fal") return model.providerModelId.toLowerCase() === KLING_O3_OMNI;
   if (model.adapter !== "replicate") return false;
   const kind = videoKind(model.providerModelId);
   return kind === "kling" || kind === "kling-omni";
 }
-function fillReplicateKlingResolution(model, req) {
-  if (isReplicateKling(model) && req.resolution === void 0) {
+function fillKlingModeResolution(model, req) {
+  if (takesKlingMode(model) && req.resolution === void 0) {
     req.resolution = klingTier(req).resolution;
   }
 }
@@ -11513,7 +11557,7 @@ var minimaxH3MaxFamily = {
 
 // src/adapters/fal-omnihuman.ts
 var OMNIHUMAN_V22 = "bytedance/omnihuman/v2";
-function plan(req, ctx) {
+function plan2(req, ctx) {
   const [image] = req.referenceImages ?? [];
   const [audio] = req.referenceAudios ?? [];
   if (image === void 0 || audio === void 0) {
@@ -11524,14 +11568,14 @@ function plan(req, ctx) {
   return { endpoint: OMNIHUMAN_V22, input, seconds: inputLengthSeconds(req, audio) };
 }
 var OMNIHUMAN_MODEL_IDS = /* @__PURE__ */ new Set([OMNIHUMAN_V22]);
-var omnihumanFamily = { ids: OMNIHUMAN_MODEL_IDS, plan };
+var omnihumanFamily = { ids: OMNIHUMAN_MODEL_IDS, plan: plan2 };
 
 // src/adapters/fal-vidu.ts
 var FAMILY2 = "fal-ai/vidu/q4";
 function viduTier(resolution) {
   return resolution === "2k" || resolution === "4k" ? upperTier(resolution) : resolution;
 }
-function plan2(req, ctx) {
+function plan3(req, ctx) {
   const images = urls(req.referenceImages);
   const audios = urls(req.referenceAudios);
   const duration = wholeSeconds(req, ctx);
@@ -11558,11 +11602,12 @@ function plan2(req, ctx) {
   return { endpoint, input, resolution, seconds: requestedSeconds(duration) };
 }
 var VIDU_MODEL_IDS = /* @__PURE__ */ new Set([FAMILY2]);
-var viduFamily = { ids: VIDU_MODEL_IDS, plan: plan2 };
+var viduFamily = { ids: VIDU_MODEL_IDS, plan: plan3 };
 
 // src/adapters/fal-video-families.ts
 var FAMILIES2 = [
   klingV3Family,
+  klingOmniFamily,
   minimaxH3MaxFamily,
   happyHorseFamily,
   viduFamily,
@@ -12039,7 +12084,7 @@ function extractVideoUrl(result, endpoint, model) {
 }
 async function generateFalVideo(req, ctx, deps) {
   const providerModelId = ctx.model.providerModelId;
-  const plan3 = planFalFamilyVideo(req, ctx);
+  const plan4 = planFalFamilyVideo(req, ctx);
   const start = (endpoint2, input, usageOf2) => startFalVideoJob(
     ctx,
     () => falClientOf(ctx, deps),
@@ -12047,7 +12092,7 @@ async function generateFalVideo(req, ctx, deps) {
     input,
     videoFileDecoder(ctx, req, endpoint2, usageOf2)
   );
-  if (plan3 !== void 0) return start(plan3.endpoint, plan3.input, planUsage(ctx, plan3));
+  if (plan4 !== void 0) return start(plan4.endpoint, plan4.input, planUsage(ctx, plan4));
   if (isKling(providerModelId)) return start(KLING_O1_ENDPOINT, klingO1Input(req));
   if (isWan(providerModelId)) {
     const { endpoint: endpoint2, input } = wanVideo(req, providerModelId.toLowerCase());
@@ -12097,9 +12142,9 @@ function videoFileDecoder(ctx, req, endpoint, usageOf2) {
     raw: payload
   });
 }
-function planUsage(ctx, plan3) {
+function planUsage(ctx, plan4) {
   return (payload) => {
-    const seconds3 = plan3.seconds(payload);
+    const seconds3 = plan4.seconds(payload);
     return {
       model: ctx.model.id,
       provider: "fal",
@@ -12107,7 +12152,7 @@ function planUsage(ctx, plan3) {
       unit: "seconds",
       quantity: seconds3,
       durationSeconds: seconds3,
-      ...plan3.resolution !== void 0 && { resolution: plan3.resolution }
+      ...plan4.resolution !== void 0 && { resolution: plan4.resolution }
     };
   };
 }
@@ -14132,6 +14177,21 @@ function imageUsage2(ctx) {
   };
 }
 var GENERATE_SECONDS = 5;
+var GENERATE_LENGTHS = [5, 10];
+function generateSeconds(req) {
+  return req.durationSeconds ?? GENERATE_SECONDS;
+}
+function checkGenerateSeconds(req, ctx) {
+  const seconds3 = generateSeconds(req);
+  if (GENERATE_LENGTHS.includes(seconds3)) return;
+  throw new ModelKitError({
+    code: "unsupported_option",
+    message: `${ctx.model.id} renders ${GENERATE_LENGTHS.join(" or ")} s clips, not ${seconds3} s`,
+    provider: "luma",
+    model: ctx.model.id,
+    retryable: false
+  });
+}
 var MAX_PLAUSIBLE_SECONDS = 60;
 function editedSeconds(req, bytes) {
   const measured = mp4DurationSeconds(bytes);
@@ -14197,7 +14257,7 @@ function videoBody(req, ctx, mode) {
       prompt,
       video: {
         ...req.resolution !== void 0 && { resolution: req.resolution },
-        duration: "5s",
+        duration: `${generateSeconds(req)}s`,
         ...req.firstFrame !== void 0 && { start_frame: imageRef(req.firstFrame) },
         ...req.lastFrame !== void 0 && { end_frame: imageRef(req.lastFrame) }
       }
@@ -14306,7 +14366,7 @@ function lumaVideoJob(ctx, req, mode, id) {
       if (mode === "generate") {
         const result2 = {
           video: { url, expiresAt: new Date(Date.now() + OUTPUT_URL_TTL_MS) },
-          usage: videoUsage4(jobCtx, req, mode, GENERATE_SECONDS),
+          usage: videoUsage4(jobCtx, req, mode, generateSeconds(req)),
           raw: { id }
         };
         return result2;
@@ -14334,6 +14394,7 @@ async function generateLumaVideo(req, ctx) {
     });
   }
   const mode = lumaVideoMode(ctx);
+  if (mode === "generate") checkGenerateSeconds(req, ctx);
   const created = await submitGeneration(ctx, videoBody(req, ctx, mode));
   return lumaVideoJob(ctx, req, mode, created.id);
 }
@@ -16545,10 +16606,10 @@ function resolveVideoOptions(model, request, onUnsupported = "error") {
     warnings
   );
   if (aspectRatio !== void 0) merged.aspectRatio = aspectRatio;
-  fillReplicateKlingResolution(model, merged);
+  fillKlingModeResolution(model, merged);
   const resolution = resolveVideoResolution(merged.resolution, model, onUnsupported, warnings);
   if (resolution !== void 0) merged.resolution = resolution;
-  if (isReplicateKling(model)) delete merged.mode;
+  if (takesKlingMode(model)) delete merged.mode;
   const duration = resolveDuration(
     merged.durationSeconds,
     c.durations,