Skip to main content
Glama

Studio: generate video

studio_generate_video
Destructive

Generate a single video clip via the Studio model catalog (call list_studio_models for keys). ASYNC — returns a jobId; poll get_job until status is succeeded/failed. SPENDS CREDITS. firstFrame/lastFrame/referenceImages/referenceVideos/referenceAudios accept URLs (import them first if needed). Video-edit models (capabilities.supportsReferenceVideos, e.g. fal-gemini-omni-flash-edit) take the source clip as referenceVideos[0] — never as a reference image. Models with capabilities.maxElements (fal-kling-4) keep characters and objects consistent: pass elements [{frontalImage, views[1–3], voiceSample?}] and name them @Element1, @Element2 … in order; an element with a voiceSample (5–30 s of that voice alone) speaks its lines in that voice. Models with capabilities.supportsShots take shots [{prompt ≤512 chars, duration}] instead of one prompt: the clip cuts between them on those seconds and runs for their sum (duration and prompt are then optional). Models with capabilities.videoModes (fal-heygen-video) take videoMode, one of the listed mode ids; without it the model's default mode runs.

Tool call template

This tool declares no required parameters. Add optional parameters from the input schema as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {},
    "name": "studio_generate_video"
  }
}

Input Schema

NameRequiredDescriptionDefault
shotsNoPer-shot prompts (capabilities.supportsShots): each ≤512 characters with whole-second durations; the clip cuts on them and runs for their sum. Each shot prompt stands alone, so repeat the style and the @tags it needs.
promptNoThe clip's prompt. Optional only when shots are given.
durationNo
elementsNoConsistent characters/objects (capabilities.maxElements): frontalImage = a front view URL, views = 1–3 more view URLs (a character sheet counts as one), voiceSample = optional URL of 5–30 s of that character's voice alone. Named @Element1 … in array order.
modelKeyNo
lastFrameNoEnd-frame image URL (models with capabilities.supportsLastFrame)
videoModeNoMode id from the model's capabilities.videoModes (e.g. "quality" on fal-heygen-video). Leave out for models without modes.
firstFrameNoStart-frame image URL
resolutionNo
aspectRatioNo
generateAudioNo
referenceAudiosNoAudio URLs (models with capabilities.supportsReferenceAudio)
referenceImagesNo
referenceVideosNoVideo URLs (models with capabilities.supportsReferenceVideos, up to maxReferenceVideos). For edit models this is the clip to edit.

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed5 schema fields changed
    • changedInput schema / properties / aspectRatio / enum
      Previous value: -[
      -  "1:1",
      -  "16:9",
      -  "9:16",
      -  "4:3",
      -  "3:4"
      -]New value: +[
      +  "1:1",
      +  "16:9",
      +  "9:16",
      +  "4:3",
      +  "3:4",
      +  "21:9"
      +]
    • addedInput schema / properties / elements
      Added value: +{
      +  "description": "Consistent characters/objects (capabilities.maxElements): frontalImage = a front view URL, views = 1–3 more view URLs (a character sheet counts as one), voiceSample = optional URL of 5–30 s of that character's voice alone. Named @Element1 … in array order.",
      +  "items": {
      +    "properties": {
      +      "frontalImage": {
      +        "maxLength": 2048,
      +        "minLength": 1,
      +        "type": "string"
      +      },
      +      "views": {
      +        "items": {
      +          "maxLength": 2048,
      +          "minLength": 1,
      +          "type": "string"
      +        },
      +        "maxItems": 3,
      +        "minItems": 1,
      +        "type": "array"
      +      },
      +      "voiceSample": {
      +        "maxLength": 2048,
      +        "minLength": 1,
      +        "type": "string"
      +      }
      +    },
      +    "required": [
      +      "frontalImage",
      +      "views"
      +    ],
      +    "type": "object"
      +  },
      +  "maxItems": 7,
      +  "type": "array"
      +}
    • addedInput schema / properties / prompt / description
      Added value: +"The clip's prompt. Optional only when shots are given."
    • addedInput schema / properties / shots
      Added value: +{
      +  "description": "Per-shot prompts (capabilities.supportsShots): each ≤512 characters with whole-second durations; the clip cuts on them and runs for their sum. Each shot prompt stands alone, so repeat the style and the @tags it needs.",
      +  "items": {
      +    "properties": {
      +      "duration": {
      +        "maximum": 30,
      +        "minimum": 1,
      +        "type": "integer"
      +      },
      +      "prompt": {
      +        "maxLength": 2000,
      +        "minLength": 1,
      +        "type": "string"
      +      }
      +    },
      +    "required": [
      +      "prompt",
      +      "duration"
      +    ],
      +    "type": "object"
      +  },
      +  "maxItems": 8,
      +  "minItems": 1,
      +  "type": "array"
      +}
    • removedInput schema / required
      Removed value: -[
      -  "prompt"
      -]
  2. Changed1 schema field changed
    • addedInput schema / properties / videoMode
      Added value: +{
      +  "description": "Mode id from the model's capabilities.videoModes (e.g. \"quality\" on fal-heygen-video). Leave out for models without modes.",
      +  "maxLength": 40,
      +  "minLength": 1,
      +  "type": "string"
      +}
  3. Changed3 schema fields changed
    • addedInput schema / properties / lastFrame
      Added value: +{
      +  "description": "End-frame image URL (models with capabilities.supportsLastFrame)",
      +  "type": "string"
      +}
    • addedInput schema / properties / referenceAudios
      Added value: +{
      +  "description": "Audio URLs (models with capabilities.supportsReferenceAudio)",
      +  "items": {
      +    "type": "string"
      +  },
      +  "maxItems": 3,
      +  "type": "array"
      +}
    • addedInput schema / properties / referenceVideos
      Added value: +{
      +  "description": "Video URLs (models with capabilities.supportsReferenceVideos, up to maxReferenceVideos). For edit models this is the clip to edit.",
      +  "items": {
      +    "type": "string"
      +  },
      +  "maxItems": 3,
      +  "type": "array"
      +}
  4. First observed

TDQS

A4.6/5.0
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations already declare destructiveHint=true and non-idempotent, but the description adds the material facts an agent needs: it is ASYNC, returns a jobId, requires polling get_job, and SPENDS CREDITS. Those are exactly the costly/irreversible behaviors the annotations cannot express.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loaded with the essential facts (model catalog, async, credits) before the capability branching. It is a dense single block and the model-capability rules could be split for scanability, but almost every clause conveys distinct, actionable information.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Covers the async lifecycle, credit cost, URL import prerequisite, and the mutually exclusive parameter modes for a 14-parameter tool with no output schema. Minor gaps remain on resolution/aspectRatio/generateAudio interactions, though those enums are largely self-explanatory.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters5/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

With only 57% schema description coverage and 14 parameters, the description carries the load well: referenceVideos[0] is the source clip for edit models, elements are addressed as @Element1 in array order with voiceSample semantics, shots replace prompt/duration, and videoMode ids come from capabilities.videoModes. This is meaning well beyond the raw schema.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

Opens with a specific verb+resource ('Generate a single video clip') and immediately scopes it to the Studio model catalog, pointing at list_studio_models for keys. An agent can tell this apart from seedance_generate_video or create_video without opening the schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives explicit operational routing: call list_studio_models for keys, then poll get_job until status is succeeded/failed, and branch on model capabilities (supportsShots, maxElements, videoModes, supportsReferenceVideos). It does not explicitly state when to prefer a different video generator, so it stops just short of full when/when-not coverage.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Try in Browser

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources