@mindstudio-ai/agent 0.1.153 → 0.1.155

package/llms.txt

@@ -275,11 +275,10 @@ Set a video's thumbnail: embedded as its cover art, and optionally as its openin
 - Output: `{ videoUrl: string, thumbnailUrl: string }`
 
 #### analyzeImage
-Analyze an image using a vision model based on a text prompt.
-- Uses the configured vision model to generate a text analysis of the image.
+Analyze one or more images with a chat model that reads images, based on a text prompt.
+- Runs on any chat model that reads images; the default is gemini-3.8-flash. `visionModelOverride` takes the same settings as a chat model's override (`model`, `temperature`, `maxResponseTokens`).
 - The prompt should describe what to look for or extract from the image.
-- Pass imageUrl for a single image, or imageUrls for multiple images analyzed together in one request.
-- Most vision models (OpenAI, Grok, Gemini) accept multiple images in one request. Ideogram describe is single-image only.
+- Pass imageUrls for one or more images analyzed together in one request (imageUrl, a single image, still works).
 - Input: `{ prompt: string, imageUrl?: string, imageUrls?: string[], visionModelOverride?: { model: string, config?: object } | { model: string, temperature: number, maxResponseTokens: number, ignorePreamble?: boolean, userMessagePreprocessor?: { dataSource?: string, messageTemplate?: string, maxResults?: number, enabled?: boolean, shouldInherit?: boolean }, preamble?: string, multiModelEnabled?: boolean, editResponseEnabled?: boolean, promptCaching?: boolean, config?: object } }`
 - Output: `{ analysis: string }`
 
@@ -463,7 +462,7 @@ Generate a 3D model using a 3D generation model.
 - From text or an image: wavespeed-tripo3d-h3.1, wavespeed-meshy7.1 (also `images`, 1–4 views), wavespeed-hunyuan3d-v3 (an `image`, plus optional `back_image_url`, `left_image_url`, `right_image_url`), wavespeed-hunyuan-3d-v3.1-rapid, wavespeed-tripo3d-p2, wavespeed-rodin-v2.5 (`images`, 1–5).
 - From an image only: trellis-2, wavespeed-meshy7 (or `images`, 1–4 views), wavespeed-tripo3d-v2.5, seed3d-2.0, wavespeed-sam-3d-objects (with optional `mask_images`).
 - From views: wavespeed-hi3d-v3.0 (`front_image`, and optional `back_image`, `left_image`, `right_image`).
-- `$mode` in the config picks a model's mode (`text-to-3d`, `image-to-3d`, or its multi-view mode). Left out, `images` picks the multi-view mode and `image` picks image-to-3d; with neither, the model's default mode. A text mode needs the prompt; in the others the prompt is optional guidance. A prompt longer than the model reads is refused, not cut.
+- `$mode` in the config picks a model's mode (`text-to-3d`, `image-to-3d`, or its multi-view mode). Left out, `images` picks the multi-view mode and `image` picks image-to-3d; with neither, the model's default mode. A text mode needs the prompt; in the others the prompt is optional guidance.
 - The output is standardized as a GLB URL plus optional FBX/OBJ/USDZ, thumbnail, texture maps, and provider task ID.
 - Input: `{ prompt?: string, intermediateAsset?: boolean, threeDModelOverride?: { model: string, config?: object } }`
 - Output: `{ prompt?: string, glbUrl: string, fbxUrl?: string, objUrl?: string, usdzUrl?: string, thumbnailUrl?: string, textureUrls?: object[], providerTaskId?: string, resolvedConfig?: object }`
@@ -479,11 +478,13 @@ Generate an HTML asset and export it as a webpage, PDF, or image
 - Output: `{ url: string }`
 
 #### generateImage
-Generate an image from a text prompt using an AI model.
+Generate an image from a text prompt, or edit one, using an AI model.
 - Prompts should be descriptive but concise (roughly 3–6 sentences).
+- The default model is seedream-4.5. Options go in `imageModelOverride.config` under the model's own names, and differ by model, so look the model up before writing the call. Images to edit or draw from go there too, as URLs: `images` on most models, `source_images` on the Gemini and GPT Image models, one `image` on qwen-image, z-image-turbo, krea-v2-turbo, ideogram-v4 and the MAI models, `source_image` on Grok Imagine, `image_url` on the Ideogram remix and FLUX Kontext models, and `reference` on krea-v2-large.
+- A model that makes and edits images by separate means takes `$mode` in its config (`text-to-image`, and `edit`, `image-edit` or `image-to-image`, as it lists). Left out, images in the config pick the edit mode, and otherwise the model makes the image from the prompt.
 - Images are hosted on the public CDN by default, or in one of the app's own file stores when the call passes `store` — a private store keeps the image private.
 - In foreground mode, the image is displayed to the user. In background mode, the URL is saved to a variable.
-- When generateVariants is true with numVariants > 1, multiple images are generated in parallel.
+- When generateVariants is true with numVariants > 1 (up to 10), the images are generated in parallel. One that fails is left out of the URLs returned, and the step fails only when every one does.
 - In direct execution, foreground mode behaves as background, and userSelect variant behavior behaves as saveAll.
 - Input: `{ prompt: string, intermediateAsset?: boolean, imageModelOverride?: { model: string, config?: object }, generateVariants?: boolean, numVariants?: number, addWatermark?: boolean }`
 - Output: `{ imageUrl: string | string[] }`
@@ -496,7 +497,7 @@ Generate a lip sync video from audio and an image or video.
 
 #### generateMusic
 Generate an audio file from provided instructions (text) using a music model.
-- `text` describes the song: genre, mood, instruments, tempo, vocals, and for Lyria its structure (`[Verse]`, `[Chorus]`, timestamps such as `[0:00-0:20] Intro`). A text longer than the model reads is refused, not cut.
+- `text` describes the song: genre, mood, instruments, tempo, vocals, and for Lyria its structure (`[Verse]`, `[Chorus]`, timestamps such as `[0:00-0:20] Intro`).
 - The default model is elevenlabs-music_v2.5. Options go in `musicModelOverride.config`, and differ by model:
 - elevenlabs-music_v2.5, elevenlabs-music_v2: `musicLengthMs`, the song's length from 3,000 to 600,000 (the model chooses one from the text otherwise, and holds a length it's given), and `forceInstrumental: true` for no vocals.
 - lyria-3.5: songs of up to about three minutes, their length steered from the text; `image`, the URL of an image for the song to take its mood and subject from. lyria-3 makes 30-second clips, lyria-3-pro songs of up to about three minutes.
@@ -542,7 +543,7 @@ the server-side plumbing.
 
 #### generateVideo
 Generate a video from a text prompt, an image or references, using an AI model.
-- Prompts should be descriptive but concise (roughly 3–6 sentences): the subject, the motion, the camera and the sound. A prompt longer than the model reads is refused, not cut.
+- Prompts should be descriptive but concise (roughly 3–6 sentences): the subject, the motion, the camera and the sound.
 - The default model is wan3.0-video. Options go in `videoModelOverride.config` under the model's own names, and differ by model, so look the model up before writing the call. Images, videos and audio go there too, as URLs: a first frame is `image` on most models, `image_url` on Veo, gemini-omni-flash, Runway's and minimax-h3-max, `source_image` on Grok Imagine, `start_frame` on ray-2 and `first_frame_image` on skyreels-v4.
 - A model that makes video more than one way takes `$mode` in its config (`text-to-video`, `image-to-video`, `reference-to-video`, `video-edit` and others it lists). Left out, a first frame or reference images in the config pick their mode, and otherwise the model's default runs.
 - Videos are hosted on the public CDN by default, or in one of the app's own file stores when the call passes `store` — a private store keeps the video private, with no public copy left behind.

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@mindstudio-ai/agent",
-  "version": "0.1.153",
+  "version": "0.1.155",
   "description": "TypeScript SDK for MindStudio direct step execution",
   "type": "module",
   "main": "./dist/index.js",