@picsart/ai-sdk 6.36.0 → 6.38.0

package/README.md

@@ -194,7 +194,7 @@ names or catalog model IDs from an arbitrary vendor request.
 
 ## Text Generation (LLMs)
 
-Claude, GPT, and Gemini text models are called with `generateText()`. Single-shot:
+Claude, GPT, Gemini, Kimi and Grok text models are called with `generateText()`. Single-shot:
 pass a prompt and optional image(s)/video, get text back. They surface in the catalog
 as `mode: 'text'`. Each vendor uses its own native workflow, so capabilities aren't lost
 — Gemini accepts video input, Claude uses the Anthropic-native messages API.
@@ -227,8 +227,9 @@ await ai.generateText(Models.Gemini3Pro, {
 catalog.find({ output: 'text' })
 ```
 
-> Thinking level maps per vendor: OpenAI → `reasoning_effort`, Gemini →
-> `thinkingConfig.thinkingLevel` (LOW/HIGH). Claude’s `claude/v1/messages`
+> Thinking level maps per vendor: OpenAI, Kimi and Grok → `reasoning_effort`, Gemini →
+> `thinkingConfig.thinkingLevel` (LOW/HIGH). Kimi K3 and Grok 4.7 always reason, so
+> they have no `off`. Claude’s `claude/v1/messages`
 > workflow exposes no thinking knob, so Claude models omit `thinking`.
 
 `generateText()` is type-narrowed to text models (`TextModelId`); calling it with an

package/index.d.ts

@@ -659,6 +659,11 @@ type ModelInputById = {
         count?: 1 | 2 | 4 | 6 | 8 | 10;
         imageUrls?: string[];
     };
+    "grok-4.7": {
+        prompt: string;
+        imageUrls?: string[];
+        thinking?: "low" | "medium" | "high";
+    };
     "grok-edit-video": {
         prompt: string;
         videoUrl: string;
@@ -1060,7 +1065,7 @@ type ModelInputById = {
         negativePrompt?: string;
         aspectRatio?: "21:9" | "16:9" | "4:3" | "1:1" | "3:4" | "9:16" | "9:21";
         resolution?: "480p" | "720p" | "1080p";
-        numFrames?: number;
+        duration?: 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20;
         fps?: number;
         sourceScale?: number;
         videoStrength?: number;
@@ -2093,7 +2098,7 @@ type ModelInputById = {
 type TypedModelId = keyof ModelInputById;
 type ModelInput<M extends TypedModelId> = ModelInputById[M];
 /** IDs of text-generation (LLM) models — narrows generateText(). */
-type TextModelId = "claude-fable-5" | "claude-fable-5-1" | "claude-haiku-4-5" | "claude-haiku-5-5" | "claude-opus-4-8" | "claude-opus-5" | "claude-sonnet-4-5" | "claude-sonnet-4-6" | "claude-sonnet-5" | "eleven-speech-to-text" | "gemini-2.5-flash" | "gemini-3-pro" | "gemini-3.5-flash-lite" | "gemini-3.6-flash" | "gemini-3.7-flash" | "gemini-3.8-flash" | "gpt-4.1-mini" | "gpt-4.1-nano" | "gpt-4o" | "gpt-4o-mini" | "gpt-5" | "gpt-5-mini" | "gpt-5.1" | "gpt-5.2" | "gpt-5.5" | "gpt-5.6-luna" | "gpt-5.6-sol" | "gpt-5.6-terra" | "gpt-6-astra" | "gpt-6-luna" | "gpt-6-sol" | "kimi-k3";
+type TextModelId = "claude-fable-5" | "claude-fable-5-1" | "claude-haiku-4-5" | "claude-haiku-5-5" | "claude-opus-4-8" | "claude-opus-5" | "claude-sonnet-4-5" | "claude-sonnet-4-6" | "claude-sonnet-5" | "eleven-speech-to-text" | "gemini-2.5-flash" | "gemini-3-pro" | "gemini-3.5-flash-lite" | "gemini-3.6-flash" | "gemini-3.7-flash" | "gemini-3.8-flash" | "gpt-4.1-mini" | "gpt-4.1-nano" | "gpt-4o" | "gpt-4o-mini" | "gpt-5" | "gpt-5-mini" | "gpt-5.1" | "gpt-5.2" | "gpt-5.5" | "gpt-5.6-luna" | "gpt-5.6-sol" | "gpt-5.6-terra" | "gpt-6-astra" | "gpt-6-luna" | "gpt-6-sol" | "grok-4.7" | "kimi-k3";
 type TextModelInputById = Pick<ModelInputById, TextModelId>;
 /** IDs of structured-output (json-mode) models — narrows run()'s result type. */
 type JsonModelId = "typesafe-evaluate";
@@ -3533,6 +3538,7 @@ declare const Models: {
     readonly GptImage2: "gpt-image-2";
     readonly GptImage25Flare: "gpt-image-2.5-flare";
     readonly GptImage25Sunburst: "gpt-image-2.5-sunburst";
+    readonly Grok47: "grok-4.7";
     readonly GrokEditVideo: "grok-edit-video";
     readonly GrokExtendVideo: "grok-extend-video";
     readonly GrokImagineImage: "grok-imagine-image";

package/index.js

@@ -2198,6 +2198,47 @@ var LTX_23_REFRAME_RESOLUTIONS = ["720p", "1080p"];
 var LTX_23_REFRAME_AR = ["16:9", "9:16", "1:1", "4:5", "5:4"];
 var LTX_23_OUTPAINT_RESOLUTIONS = ["480p", "720p", "1080p"];
 var LTX_23_OUTPAINT_AR = ["21:9", "16:9", "4:3", "1:1", "3:4", "9:16", "9:21"];
+var LTX_23_OUTPAINT_DURATIONS = [
+  1,
+  2,
+  3,
+  4,
+  5,
+  6,
+  7,
+  8,
+  9,
+  10,
+  11,
+  12,
+  13,
+  14,
+  15,
+  16,
+  17,
+  18,
+  19,
+  20
+];
+var LTX_23_OUTPAINT_MIN_FRAMES = 9;
+var LTX_23_OUTPAINT_MAX_FRAMES = 481;
+var LTX_23_OUTPAINT_FPS_MAX = 60;
+var ltxOutpaintDurationConstraints = [];
+for (let fps = 1; fps <= LTX_23_OUTPAINT_FPS_MAX; fps++) {
+  const allowed = LTX_23_OUTPAINT_DURATIONS.filter(
+    (d) => d * fps >= LTX_23_OUTPAINT_MIN_FRAMES && d * fps <= LTX_23_OUTPAINT_MAX_FRAMES
+  );
+  if (allowed.length === LTX_23_OUTPAINT_DURATIONS.length) continue;
+  ltxOutpaintDurationConstraints.push({
+    when: { fps: { is: fps } },
+    then: {
+      duration: {
+        allowed,
+        reason: `At ${fps} fps the clip must run ${allowed[0]}\u2013${allowed[allowed.length - 1]}s (the ${LTX_23_OUTPAINT_MIN_FRAMES}\u2013${LTX_23_OUTPAINT_MAX_FRAMES}-frame window).`
+      }
+    }
+  });
+}
 var LTX_25_AR = ["16:9", "9:16"];
 var LTX_25_PRO_RESOLUTIONS = ["720p", "1080p"];
 var LTX_25_FAST_RESOLUTIONS = ["720p", "1080p", "1440p", "2160p"];
@@ -2543,8 +2584,11 @@ var { MODELS: MODELS2 } = defineModels("ltx", [
       ...params.negativePrompt(),
       ...params.aspectRatio(LTX_23_OUTPAINT_AR),
       ...params.resolution(LTX_23_OUTPAINT_RESOLUTIONS, "720p"),
-      ...p.range("numFrames", 9, 481, 121, { label: "Frames" }),
-      ...p.range("fps", 1, 60, 24, { label: "FPS" }),
+      // Duration is an enum; fps stays a slider. The per-fps constraint below
+      // trims the duration options, and the builder derives the wire's num_frames
+      // (clamped) from duration × fps.
+      ...params.duration(LTX_23_OUTPAINT_DURATIONS, 6),
+      ...p.range("fps", 1, 60, 24, { step: 1, label: "FPS" }),
       // 1.0 fits the source as large as the canvas allows; lower values zoom it
       // out and leave a wider generated border, which the vendor calls less stable.
       ...p.range("sourceScale", 0.25, 1, 1, { step: 0.05, label: "Source Scale" }),
@@ -2559,7 +2603,10 @@ var { MODELS: MODELS2 } = defineModels("ltx", [
       ...p.boolean("enableSafetyChecker", true, "Safety Checker"),
       // Default-less on purpose: unset means the vendor picks a random seed.
       ...params.seed()
-    }
+    },
+    // Raising fps packs more frames per second, so it shortens the longest clip
+    // that fits the vendor's 481-frame ceiling — hence the duration options trim.
+    constraints: ltxOutpaintDurationConstraints
   },
   // ── LTX 2.5 ──────────────────────────────────────────────────────
   // One entry per tier: text-to-video on `workflow`, image-to-video on
@@ -10579,7 +10626,27 @@ var { MODELS: KIMI_LLM } = defineModels("kimi", [
     }
   }
 ]);
-var MODELS35 = [...ANTHROPIC, ...OPENAI_LLM, ...GEMINI_LLM, ...KIMI_LLM];
+var grokThinkingParam = () => p.enum("thinking", ["low", "medium", "high"], "high", { label: "Thinking" });
+var { MODELS: GROK_LLM } = defineModels("grok", [
+  {
+    id: "grok-4.7",
+    name: "Grok 4.7",
+    workflow: "chat-completions",
+    addedAt: "2026-10-14",
+    estimatedTime: 10,
+    mode: "text",
+    inputType: "i2t",
+    badge: ["new"],
+    description: "xAI\u2019s flagship Grok model \u2014 always-on reasoning with a 500K-token context.",
+    features: [feat("Vision", "input"), feat("Thinking", "characteristic")],
+    paramConfig: {
+      ...params.prompt(),
+      ...params.imageInput(8, "Images"),
+      ...grokThinkingParam()
+    }
+  }
+]);
+var MODELS35 = [...ANTHROPIC, ...OPENAI_LLM, ...GEMINI_LLM, ...KIMI_LLM, ...GROK_LLM];
 
 // src/vendors/catalog/llm.payloads.ts
 var CLAUDE_MAX_TOKENS = 8192;
@@ -10663,7 +10730,8 @@ registerPayloads(MODELS35, {
   "gemini-3.6-flash": buildOpenAiPayload("gemini-3.6-flash"),
   "gemini-3.5-flash-lite": buildOpenAiPayload("gemini-3.5-flash-lite"),
   "gemini-2.5-flash": buildOpenAiPayload("gemini-2.5-flash"),
-  "kimi-k3": buildOpenAiPayload("kimi-k3")
+  "kimi-k3": buildOpenAiPayload("kimi-k3"),
+  "grok-4.7": buildOpenAiPayload("grok-4.7")
 });
 
 // src/vendors/catalog/captionsai.ts
@@ -11022,25 +11090,34 @@ var buildLtx23ReframePayload = (input) => ({
   ...input.resolution ? { resolution: input.resolution } : {},
   ...input.aspectRatio ? { aspect_ratio: input.aspectRatio } : {}
 });
-var buildLtx23OutpaintPayload = (input) => ({
-  prompt: input.prompt,
-  video_url: input.videoUrl,
-  ...input.negativePrompt ? { negative_prompt: input.negativePrompt } : {},
-  ...input.aspectRatio ? { aspect_ratio: input.aspectRatio } : {},
-  ...input.resolution ? { output_resolution: input.resolution } : {},
-  ...input.numFrames != null ? { num_frames: input.numFrames } : {},
-  ...input.fps != null ? { frames_per_second: input.fps } : {},
-  ...input.sourceScale != null ? { source_scale: input.sourceScale } : {},
-  ...input.videoStrength != null ? { video_strength: input.videoStrength } : {},
-  ...input.cfgScale != null ? { guidance_scale: input.cfgScale } : {},
-  ...input.numInferenceSteps != null ? { num_inference_steps: input.numInferenceSteps } : {},
-  ...input.videoQuality ? { video_quality: input.videoQuality } : {},
-  ...input.videoWriteMode ? { video_write_mode: input.videoWriteMode } : {},
-  ...input.seed != null ? { seed: input.seed } : {},
-  enable_prompt_expansion: input.enhancePrompt ?? true,
-  generate_audio: input.generateAudio ?? true,
-  enable_safety_checker: input.enableSafetyChecker ?? true
-});
+var OUTPAINT_NUM_FRAMES_MIN = 9;
+var OUTPAINT_NUM_FRAMES_MAX = 481;
+var buildLtx23OutpaintPayload = (input) => {
+  const fps = input.fps ?? 24;
+  const numFrames = input.duration != null ? Math.min(
+    OUTPAINT_NUM_FRAMES_MAX,
+    Math.max(OUTPAINT_NUM_FRAMES_MIN, Math.round(input.duration * fps))
+  ) : void 0;
+  return {
+    prompt: input.prompt,
+    video_url: input.videoUrl,
+    ...input.negativePrompt ? { negative_prompt: input.negativePrompt } : {},
+    ...input.aspectRatio ? { aspect_ratio: input.aspectRatio } : {},
+    ...input.resolution ? { output_resolution: input.resolution } : {},
+    ...numFrames != null ? { num_frames: numFrames } : {},
+    ...input.fps != null ? { frames_per_second: input.fps } : {},
+    ...input.sourceScale != null ? { source_scale: input.sourceScale } : {},
+    ...input.videoStrength != null ? { video_strength: input.videoStrength } : {},
+    ...input.cfgScale != null ? { guidance_scale: input.cfgScale } : {},
+    ...input.numInferenceSteps != null ? { num_inference_steps: input.numInferenceSteps } : {},
+    ...input.videoQuality ? { video_quality: input.videoQuality } : {},
+    ...input.videoWriteMode ? { video_write_mode: input.videoWriteMode } : {},
+    ...input.seed != null ? { seed: input.seed } : {},
+    enable_prompt_expansion: input.enhancePrompt ?? true,
+    generate_audio: input.generateAudio ?? true,
+    enable_safety_checker: input.enableSafetyChecker ?? true
+  };
+};
 registerPayloads(MODELS2, {
   "ltx-v2.5-pro": buildLtx25ProT2VPayload,
   "ltx-v2.5-fast": buildLtx25FastT2VPayload,
@@ -14699,6 +14776,7 @@ var GptImage15 = "gpt-image-1.5";
 var GptImage2 = "gpt-image-2";
 var GptImage25Flare = "gpt-image-2.5-flare";
 var GptImage25Sunburst = "gpt-image-2.5-sunburst";
+var Grok47 = "grok-4.7";
 var GrokEditVideo = "grok-edit-video";
 var GrokExtendVideo = "grok-extend-video";
 var GrokImagineImage = "grok-imagine-image";
@@ -14962,6 +15040,7 @@ var Models = {
   GptImage2,
   GptImage25Flare,
   GptImage25Sunburst,
+  Grok47,
   GrokEditVideo,
   GrokExtendVideo,
   GrokImagineImage,

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@picsart/ai-sdk",
-  "version": "6.36.0",
+  "version": "6.38.0",
   "type": "module",
   "description": "Type-safe SDK for 100+ AI models — image, video, audio, and text generation with Picsart",
   "license": "MIT",