Product updates

​

Motion Graphics Videos in the Pictory MCP Server

• Choose your video style first. Ask for a standard video or motion graphics. Claude helps you settle the direction before creating the scenes.
• Create in Claude, continue in Pictory 3.0. Every motion graphics video is saved as an editable project in your connected Pictory account. Open the project link to refine it in the Motion Studio editor in Pictory 3.0 without rebuilding your scenes.
• Choose the right Brand Kit 2.0. Select a kit from Pictory 3.0 or choose None. Bring your colors, fonts, logo, tone of voice, and supplied intro and outro into the video.
• Direct the sound and movement. Choose narration or no voice-over, music throughout or scene by scene, and transitions that suit the story. Finalize narration, narrator, music, and scene count before creation; later chat revisions refine the visual design.
• Use the media you want. Choose stock media, AI-generated images and clips, both, or graphics alone. Select your AI models, and choose an alternative if AI Credits are insufficient.
• Review before you finish. Preview the video, request visual revisions, and receive share and download links when the finished video is ready.

​

Music Chosen Scene by Scene in the Pictory MCP Server

• A new question. If you want music, Claude asks whether one track plays under the whole video or a mix of tracks is chosen scene by scene. When Pictory designs it, one track plays under the whole video, as before.
• A track per part of the story. For a mix, Claude waits until the scene list is agreed, groups neighboring scenes that share a mood, searches Pictory’s music library separately for each group, and offers a different track for each. A track plays on across its group without restarting.
• The scenes in between. If some scenes are left without a track, Claude asks whether they play one track underneath or have no music. Every track in a mix is chosen from the library; there is no automatic pick for a scene.
• Levels. Each track sits at a low level under narration, or under an AI clip’s own sound, unless you ask for another level.
• Editing a preview. The music of a mix is one soundtrack on the preview: in place, you can make all of it quieter, or replace it with one track that then plays under every scene, including any that had no music. A different track on some scenes, or music louder than designed, means a fresh preview, which reuses the narration already voiced.

​

Scene-Level Music: A Different Track and Volume for Each Scene

• musicUrl. The music for a scene: a public link to an audio file, or an audioUrl from Search Music Tracks. Tracks from Search Music Tracks are licensed for you.
• volume. How loud the scene’s music is, from 0 to 1. Use it with musicUrl. If left out, the scene uses the video’s music volume.
• enabled. No longer required when you set musicUrl. Set it to false for no music in a scene, as before.
• One soundtrack. All the scene music is joined into one soundtrack:
  • Scenes next to each other with the same musicUrl keep the music playing without starting over.
  • The video’s music picks up where it would have been.
  • Short tracks repeat, and changes between tracks fade.
• Music on a few scenes only. Set the top-level backgroundMusic.enabled to false, and scenes without a musicUrl are silent.
• Previews and saved projects. The storyboard has one music element that plays the whole soundtrack. A project saved with saveProject: true keeps it as one custom music track.
• YouTube clearance. If a video uses more than one library track, youtubeClearanceCode covers the one that plays the longest.
• New job error codes:
  • MUSIC_URL_UNREADABLE: a musicUrl could not be read as audio.
  • MUSIC_LICENSE_FAILED: a library track could not be licensed.
  • SCENE_MUSIC_BUILD_FAILED: the soundtrack could not be built.

​

AI-Generated Media in the Pictory MCP Server, and Scenes Carried by an AI Background

• AI-generated media, step by step. Claude asks whether to use stock or AI-generated media, then images, videos or let AI decide, whether the clips have sound, which model, and, for two or more AI scenes, continuity and a subject that must look the same.
• New tool: list_ai_models. Lists the models that fit your answers and the video’s shape, cheapest first, with the AI Credits for one clip or image. The server now has fourteen tools.
• Credits before any preview. estimate_layout, preview_video_from_text and create_video_from_text return the AI Credits a design draws, with cheaper options such as another model, no sound on narrated clips, or stock media. Claude goes ahead only when you agree.
• Scenes carried by their clip. A person speaking in the clip, or its own sound such as a cafe or waves, can carry a scene with no narration. Claude does not ask about a voice or subtitles for those scenes, and any music plays quietly under the clip’s sound.
• Data stays on top. Generated media shows the scene only; numbers, labels and charts are drawn as elements over it, and Pictory adds where the subject should sit so they land on plain background.
• Preview or render, in either mode. Claude now offers a preview first or the finished video straight away in Design it together too, and says what each means.
• Timing. Each AI-generated clip adds 3–4 minutes to a preview or render, because clips are generated one after another.
• Pictory designs it. An AI-generated background you name is generated for every sentence of that scene’s narration, and the result now reports the estimated AI Credits.

• No content source needed. A scene with no story, brief, blogUrl, pptUrl, audioUrl or videoUrl is accepted when background.aiVisual is set. Previously only a visualUrl or a color background allowed this.
• Prompt required. Such a scene must set background.aiVisual.prompt, because there is no story to write one from.
• Sound plays on these scenes too. An AI video clip generated with audio: true already played its sound in a narrated scene; it now does in a scene with no story as well, such as dialogue spoken in the clip or ambience, unless settings.mute is true.

​

AI Studio and Storyboard AI Visuals: Audio, New Models, Multi-Reference Editing, Frame Transitions

• Audio. POST /v1/aistudio/videos accepts audio: true to generate synchronized sound, effects, and speech with the video. Audio is a separate per-second rate per model. omni-flash always produces audio; audio: false is rejected there.
• New video models. pixverse6 (up to 15s, eight aspect ratios including 21:9, extension, transitions) and omni-flash (10s, audio always on). pixverse5.6 is the documented budget option. The default model is pixverse5.6; requests that still send the retired pixverse5.5 are served and billed as pixverse5.6.
• New image models. nanobanana2 (8 credits) and nanobanana-lite2 (4 credits, edits served by nanobanana2 at the lite rate).
• Multi-reference image editing. POST /v1/aistudio/images accepts referenceImageUrls with up to four images. The singular referenceImageUrl keeps working.
• Frame transitions. endFrameImageUrl, alongside firstFrameImageUrl, generates a clip that starts on one image and ends on another (pixverse6, pixverse5.6).
• Reference limits per model. Video referenceImageUrls now allows up to four images on Pixverse models; Veo keeps the existing limit of three, and Omni Flash takes one. On veo3.1 and veo3.1_fast, reference images require an 8s duration; other values are rejected before credits are reserved.
• Highest quality by default. Generated videos are delivered at the highest resolution the model supports, 1080p on Pixverse and Veo (720p for pixverse6 at 15s, Veo clips shorter than 8s, and Veo extensions; omni-flash tops out at 720p). Credits are unchanged, and the capabilities endpoint reports maxResolution per model. Send resolution: "720p" for a lighter file.
• Listing fields. Videos carry audio and endFrameImageUrl; images carry referenceImageUrls. Both carry enhancedPrompt when prompt enhancement ran, and completed jobs return it too.
• Model capabilities endpoint. GET /v1/aistudio/models lists every model with its accepted ratios, durations, reference limits, audio support, and features, generated from the same rules that validate requests.
• Training presets. trainingPreset (sales-training, product-training, soft-skills-training) steers the prompt enhancer toward training content on both endpoints, and always turns enhancement on, even when enhancePrompt is false.
• Storyboard AI visuals. scenes[].background.aiVisual and media element aiVisual on the storyboard and render endpoints get the same models and options: pixverse6, pixverse5.6, omni-flash, nanobanana2, nanobanana-lite2, nanobanana-pro, nova-canvas; referenceImageUrls limits per model instead of 2 on every video model (4 on Pixverse, 3 on Veo, 1 on omni-flash), and on image scenes for the first time (up to 4, 1 on nova-canvas); audio (off by default, with-audio rate); endFrameImageUrl transitions; trainingPreset (always turns prompt enhancement on); enhancePrompt (off by default, so written prompts are used as written); a clip generated with audio plays it unless settings.mute is true; and resolution (1080p by default).

​

Pictory MCP Server for Claude and Claude Code

• Two design modes: let Pictory’s AI Storyboard design every scene from a brief, a script or a blog URL, or design each scene together with Claude and edit the preview in place before rendering.
• Thirteen tools: create and preview videos, render an approved preview, inspect and edit a preview’s elements, estimate a layout, search stock media and music, list voices, brand kits and text styles, and check your quota.
• Your account, your plan: there is no API key to manage. Sign-in is your Pictory account, and video minutes, premium voice minutes and AI Credits are drawn from its API subscription.

​

YouTube Content ID Clearance Code for Licensed Background Music

• New fields on GET /v1/jobs/{jobId} for completed render jobs: youtubeClearanceCode and youtubeClearanceExpiresAt. Enter the code in YouTube Studio to release the claim on that video.
• Present only when needed. Jobs with no background music, or with your own musicUrl, are unchanged. The fields also appear in the job completion webhook payload.
• One code per video, valid for 7 days. Polling the job again returns the same code until it expires, after which a fresh code is issued. Each code clears a single video upload.
• Works for existing jobs. Any completed job with library music returns a code on its next status call.

​

AI Storyboard: Let AI Design Every Scene

• Script mode: send story or blogUrl scenes and the AI designs around your narration, word for word. Story scenes are split one designed scene per sentence by default.
• Brief mode: send a single brief scene and the AI writes the narration and designs the storyboard in one pass. A render job returns the generated narration in txtFile, srtFile, and vttFile; a preview job returns it in storyboard.scenes[].story.
• Text-to-video only: templates, smart layouts, variables, the 4:5 ratio, and presentation, audio, and video sources are rejected with an explicit message. The design step adds 1–3 minutes to the job.

​

Slide Animation Controls: Model Selection and Per-Slide Targeting

{
  "pptUrl": "https://storage.example.com/deck.pptx",
  "pptSlideAnimation": {
    "enabled": true,
    "model": "veo3.1_fast",
    "slides": [1, 4, 9]
  }
}

• enabled: set to true to animate slides.
• model: choose the video model, one of pixverse6, veo3.1_fast, or veo3.1. Defaults to pixverse6 when omitted, which does not consume AI credits. The veo3.1 and veo3.1_fast models consume AI credits and reproduce small slide text more reliably. Clips are capped at 10 seconds.
• slides: animate only the slide positions you list, starting at 1. Omit or pass an empty array to animate every eligible slide.

​

Media Elements: Video and Image Overlays at Scene Creation

• Your own media: set visualUrl to any publicly reachable video or image file. Duration and proportions are detected automatically.
• Stock library search: set searchFilter.query to a plain-language description and Pictory finds a matching stock visual. Respects the request-level language.
• AI generation: set aiVisual with a prompt and model (same models and AI Credit rates as background AI visuals).

• Anchor presets (top-left through bottom-right) or exact top/left percentage offsets.
• Explicit width is used as given: "100%" with top: "0%", left: "0%" fills the whole frame.
• No width? The element sizes itself automatically and shrinks to stay fully on screen, based on its position and the media’s proportions.
• Video playback settings: loop and mute, both defaulting to true so overlays never compete with your narration.
• Per-element color overlay: tint an individual element with colorOverlay, independent of the scene background.

​

Background Color Overlay: Tint Any Scene Background

• Two fields: color (RGB, RGBA, or hex) and opacity (0–1, default 0.4).
• Works everywhere: auto-selected visuals, stock search, direct URLs, AI-generated visuals, solid color backgrounds, backgroundBrolls[] items (a B-roll’s own overlay takes precedence while it is on screen), and video/image media elements.
• Scenes with a colorOverlay are routed to the v3 storyboard engine automatically.

​

Scene Elements: Shape and Text Overlays at Scene Creation

• Shapes: 3 basic shapes (rectangle, circle, line) plus 130+ library shapes across families like badges, arrows, stars, quotes, doodles, blobs, hearts, and speech bubbles.
• Text: Three decorative presets (heading, subheading, body) with optional style overrides for font, color, alignment, case, and decorations.
• Positioning: Nine anchor regions (top-left through bottom-right) with percentage-based width, top, and left offsets.

​

PowerPoint Animated Slides

• New guide: PowerPoint to Video with Animated Slides. Covers animatePPT, feature combinations, and best practices for when to apply animations.
• Existing PowerPoint AI Voiceover and Speaker Notes guides now document animatePPT.
• The field is also documented in the Create Storyboard Preview API reference.

​

Multilingual PowerPoint Narration: Translate Decks at Render Time

• 14 supported languages: Chinese (zh), Dutch (nl), English (en), French (fr), German (de), Hindi (hi), Italian (it), Japanese (ja), Korean (ko), Marathi (mr), Portuguese (pt), Russian (ru), Spanish (es), and Tamil (ta).
• One deck → many language outputs: ship the same source PPT to multiple international audiences without maintaining translated copies.
• Works with both narration sources: whether the AI is reading the slide text or the speaker notes (useSpeakerNotes: true), the output is translated into the target language.
• Combines with animatePPT: add slide animations on top of multilingual narration in the same request.
• Documented voice per language: each language has a default male STD voice you can use straight away (e.g., Martin for English, Gabriel for French, Wilbur for German). See the End-to-End Recipes guide for the full language → voice mapping.

​

LLM-Friendly Documentation

• OpenAPI 3.1 specification: Available at docs.pictory.ai/openapi.json. Covers the five most-used endpoints (Create Storyboard Preview, Render Storyboard Video, Render Project, Get Job by ID, Get Video Brands) with 34 component schemas. Use it for SDK code generation, Postman or Insomnia import, or as a structured tool schema for LLMs.
• End-to-End Recipes guide: Ten complete, ready-to-run JSON payloads for the most common use cases, including text-to-video, multilingual narration, PowerPoint, avatar walk-throughs, template-based renders, brand kits, webhooks, previews, and job polling.
• Use Pictory with an LLM guide: Recommended system prompt, example natural-language to API-call mappings, and Pictory MCP server pointers.
• AI Tools setup guides: Project rules and configuration templates for Claude Code, Cursor, and Windsurf. Each guide teaches the tool the Pictory API conventions, including the raw-key authentication pattern and mutually-exclusive field constraints.
• llms.txt and llms-full.txt: Mintlify auto-generates these at docs.pictory.ai/llms.txt and docs.pictory.ai/llms-full.txt. Hand either URL to any LLM to make it Pictory-aware in one shot.

​

AI Credits: Unified Currency for Generative AI Features

​

Account and Usage API

• Retrieve real-time usage for video minutes, transcription minutes, ElevenLabs voice-over minutes, and AI credits.
• Each metric shows used and limit values for the current billing term.
• Use this to display usage summaries, gate features based on remaining quota, or verify capacity before launching jobs.

• Retrieve a per-transaction ledger of AI credit activity with date, amount, source, and direction (credit or debit).
• Filter by year and month to query specific billing periods.
• Transaction sources include AI Avatar, Text-to-Image, AI Video, Add-on Purchase, and Promotional Credits.
• History available for the last 12 months.

​

ElevenLabs Voice Auto-Discovery

• No manual voice registration: the Add Voiceover Track endpoint is no longer needed for ElevenLabs voices.
• Pass any ElevenLabs voice ID as the speaker field (e.g., "pNInz6obpgDQGcFmaJgB").
• Automatic library addition: once discovered, the voice is available by name or track ID in future requests.
• Premium voice settings: fine-tune stability, similarity boost, style, and model selection.

​

AI Visuals: Visual Continuity, Reference Images, and Creative Direction

• Enable visualContinuity to create seamless transitions between consecutive AI-generated scenes.
• Works with both video clips and images.
• The system uses the output of each scene as a reference for the next, producing a cohesive visual flow.
• Continuity applies within the same story and across consecutive stories when enabled.

• Use firstFrameImageUrl to control the starting frame of an AI-generated video clip.
• The AI model generates a video that begins from your provided image and transitions into the motion described by the prompt.

• Use referenceImageUrl to guide the style and composition of AI-generated images with a reference image.
• Influence the color palette, lighting, and visual tone while generating new content from your prompt.

• Use referenceImageUrls to provide 1–2 reference images that guide the style of AI-generated video clips.
• When using veo3.1 or veo3.1_fast with reference images, the video duration is automatically set to "8s".

• When a story is split into multiple scenes, the prompt field acts as a creative direction for the entire video.
• The system uses your creative direction to guide the auto-generated prompts for each individual scene.
• Recommended structure: [Action/Movement] + [Scene/Environment] + [Camera Technique] + [Visual Style].

• Job responses now include aiCreditsUsed when AI visuals are generated, reporting total credits consumed across all scenes.

​

AI Studio: AI Image and Video Generation

• Generate images from text prompts with your choice of AI model and visual style.
• Use a reference image to create variations, replace subjects, or apply transformations.
• Choose from models including seedream3.0, flux-schnell, nanobanana, and nanobanana-pro.
• Apply visual styles such as photorealistic, artistic, cartoon, minimalist, vintage, and futuristic.

• Generate videos from text prompts with configurable duration and aspect ratio.
• Start a video from a specific image using the first frame input for precise visual control.
• Extend an existing video with new AI-generated content that continues from the original.
• Provide up to three reference images to guide subject appearance and scene composition.
• Choose from models including veo3.1, veo3.1_fast, and pixverse5.5.

• Retrieve paginated lists of all your generated images and videos.
• Results are sorted by creation date with the most recent items first.

• Every completed job reports the number of AI credits consumed.
• Image generation is charged per image, and video generation is charged per second of output.
• See the rate card in each model’s documentation for exact pricing.

​

Dynamic Captions with Word-Level Timing

• Set maxSubtitleLines to a value from 1 to 4 to control how many lines of subtitles are displayed at a time.
• Word-level timing is applied by default when dynamic captions are active.