Skip to main content
Glama
Venice MCP Server

Venice MCP Server

Official
by veniceai

Venice Video Queue

venice_video_generate

Queue AI video generation from a prompt and model id, including text-to-video, image-to-video, and video-to-video. Returns a queue id to poll for the finished render.

Instructions

Queue a video generation. Supports Veo 3.1, Kling, Wan, Seedance, MiniMax, PixVerse, Runway, LTX, Grok Imagine, and others. Pick a specific id like "veo3.1-fast-text-to-video", "veo3.1-fast-image-to-video", "kling-2.6-pro-text-to-video", "wan-2.6-text-to-video", "seedance-2-5-reference-to-video-basic" etc. Uncensored: NSFW prompts allowed where the model permits. Supports x402 wallet auth (no Venice account needed) and API key. Returns { model, queue_id }; poll with venice_video_status. NOTE: 'duration' is a model-specific string enum like '4s' / '6s' / '8s'; inspect it with venice_model_details. Current public Seedance models may reject detectable persons outright. Consent flags are defensive compatibility support, not a policy bypass; set them only after showing a returned policy to the user and receiving explicit confirmation.

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "prompt": "<prompt>",
      "model": "<model>",
      "duration": "<duration>"
    },
    "name": "venice_video_generate"
  }
}

Input Schema

NameRequiredDescriptionDefault
seedNo
audioNoEnable or disable audio generation for models that support it. Defaults to true.
modelYesRequired. Full model id, e.g. "veo3.1-fast-text-to-video".
promptYes
consentsNoDefensive compatibility support for a returned Seedance needs_consent flow. Current public models may reject detectable persons outright. All three flags must be true after explicit user confirmation and cannot bypass content policy.
durationYesDuration as model-specific string enum, e.g. "4s", "6s", "8s". See venice_model_details.
elementsNoFor Kling O3 R2V and similar: up to 4 character/object elements. Reference in prompt as @Element1, @Element2, etc.
audio_urlNoFor models that support audio input: background music. URL or data URL. Supported: WAV, MP3. Max 30s, 15MB.
image_urlNoFor image-to-video models: starting frame. URL or data URL.
video_urlNoFor video-to-video and upscale models (e.g. wan-2-7-video-to-video, topaz-video-upscale): input video. URL or data URL. Supported: MP4, MOV, WebM.
resolutionNoOutput resolution, e.g. "720p", "1080p", "4k". Model-specific; see venice_model_details.
aspect_ratioNoOutput aspect ratio, e.g. "16:9", "9:16", "1:1", "4:5", "9:21". Model-specific; see venice_model_details.
end_image_urlNoFor models that support end frames or transitions. URL or data URL.
upscale_factorNoFor upscale models only: 1 = quality enhance, 2 = double resolution, 4 = quadruple.
negative_promptNoNegative prompt (what to avoid). Supported by Seedance and other models.
scene_image_urlsNoFor models with advanced element support: up to 4 scene reference images. Reference in prompt as @Image1, @Image2, etc.
video_referencesNoFor models with per_reference_audio, e.g. wan-2-7-reference-to-video: reference clips, up to the model's reference-video limit. Each has an optional ref_name to cite in the prompt and an optional audio_url (WAV or MP3) with that character's voice. Each URL may be a data URL.
camera_trajectoryNoFor minimax-h3-max-multi-angle only: 2–12 camera keyframes with strictly increasing time from 0 to 1, azimuth in degrees (signed, at most 32 full turns of travel in total), elevation from -90 to 90 degrees, and distance relative to the starting camera (1 = unchanged). Requires image_url; the output follows its aspect ratio. Omit to let the model choose the camera path.
reference_audio_urlsNoFor Seedance 2.0 R2V and similar: up to 3 reference audio clips for vocal timbre, narration, or sound effects. Per-clip 2–15s, WAV/MP3; aggregate ≤15s. Must be paired with at least one reference image or video. Each a URL or data URL.
reference_image_urlsNoFor models with reference image support: up to 9 images for character/style consistency. Each a URL or data URL.
reference_video_urlsNoFor Seedance 2.0 R2V and similar: up to 3 reference video clips to inherit subject motion, camera movement, and style. Per-clip 2–15s, MP4/MOV, ≤50MB; aggregate ≤15s. Each a URL or data URL.
omni_reference_task_typeNoFor seedance-2-5-reference-to-video-basic only: what to do with reference_video_urls. "edit" changes the referenced clip and "extend" continues it; both need reference_video_urls and a prompt that asks for the same thing. Omit to let Venice infer it from the prompt.

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed4 schema fields changedv0.3.2
    • addedInput schema / properties / camera_trajectory
      Added value: +{
      +  "description": "For minimax-h3-max-multi-angle only: 2–12 camera keyframes with strictly increasing time from 0 to 1, azimuth in degrees (signed, at most 32 full turns of travel in total), elevation from -90 to 90 degrees, and distance relative to the starting camera (1 = unchanged). Requires image_url; the output follows its aspect ratio. Omit to let the model choose the camera path.",
      +  "items": {
      +    "additionalProperties": false,
      +    "properties": {
      +      "azimuth": {
      +        "type": "number"
      +      },
      +      "distance": {
      +        "exclusiveMinimum": 0,
      +        "type": "number"
      +      },
      +      "elevation": {
      +        "maximum": 90,
      +        "minimum": -90,
      +        "type": "number"
      +      },
      +      "time": {
      +        "maximum": 1,
      +        "minimum": 0,
      +        "type": "number"
      +      }
      +    },
      +    "required": [
      +      "time",
      +      "azimuth",
      +      "elevation",
      +      "distance"
      +    ],
      +    "type": "object"
      +  },
      +  "maxItems": 12,
      +  "minItems": 2,
      +  "type": "array"
      +}
    • addedInput schema / properties / omni_reference_task_type
      Added value: +{
      +  "description": "For seedance-2-5-reference-to-video-basic only: what to do with reference_video_urls. \"edit\" changes the referenced clip and \"extend\" continues it; both need reference_video_urls and a prompt that asks for the same thing. Omit to let Venice infer it from the prompt.",
      +  "enum": [
      +    "auto",
      +    "reference",
      +    "edit",
      +    "extend"
      +  ],
      +  "type": "string"
      +}
    • addedInput schema / properties / video_references
      Added value: +{
      +  "description": "For models with per_reference_audio, e.g. wan-2-7-reference-to-video: reference clips, up to the model's reference-video limit. Each has an optional ref_name to cite in the prompt and an optional audio_url (WAV or MP3) with that character's voice. Each URL may be a data URL.",
      +  "items": {
      +    "additionalProperties": false,
      +    "properties": {
      +      "audio_url": {
      +        "format": "uri",
      +        "type": "string"
      +      },
      +      "ref_name": {
      +        "type": "string"
      +      },
      +      "video_url": {
      +        "format": "uri",
      +        "type": "string"
      +      }
      +    },
      +    "required": [
      +      "video_url"
      +    ],
      +    "type": "object"
      +  },
      +  "type": "array"
      +}
    • changedInput schema / properties / video_url / description
      Previous value: -"For video-to-video models (e.g. seedance-2-0-r2v): input video. URL or data URL. Supported: MP4, MOV, WebM."New value: +"For video-to-video and upscale models (e.g. wan-2-7-video-to-video, topaz-video-upscale): input video. URL or data URL. Supported: MP4, MOV, WebM."
  2. Changed6 schema fields changedv0.3.0
    • addedInput schema / properties / aspect_ratio / description
      Added value: +"Output aspect ratio, e.g. \"16:9\", \"9:16\", \"1:1\", \"4:5\", \"9:21\". Model-specific; see venice_model_details."
    • removedInput schema / properties / aspect_ratio / enum
      Removed value: -[
      -  "16:9",
      -  "9:16",
      -  "1:1",
      -  "2:3",
      -  "3:2",
      -  "3:4",
      -  "4:3",
      -  "21:9"
      -]
    • addedInput schema / properties / consents
      Added value: +{
      +  "additionalProperties": false,
      +  "description": "Defensive compatibility support for a returned Seedance needs_consent flow. Current public models may reject detectable persons outright. All three flags must be true after explicit user confirmation and cannot bypass content policy.",
      +  "properties": {
      +    "seedance": {
      +      "additionalProperties": false,
      +      "properties": {
      +        "confirmed_legal_right": {
      +          "const": true,
      +          "description": "Set true only after the user confirms legal rights/consent for every depicted person and all submitted media.",
      +          "type": "boolean"
      +        },
      +        "confirmed_screening_acknowledged": {
      +          "const": true,
      +          "description": "Set true only after the user acknowledges automated media screening.",
      +          "type": "boolean"
      +        },
      +        "confirmed_terms_and_privacy": {
      +          "const": true,
      +          "description": "Set true only after the user confirms the returned policy text and Venice terms/privacy.",
      +          "type": "boolean"
      +        }
      +      },
      +      "required": [
      +        "confirmed_terms_and_privacy",
      +        "confirmed_legal_right",
      +        "confirmed_screening_acknowledged"
      +      ],
      +      "type": "object"
      +    }
      +  },
      +  "required": [
      +    "seedance"
      +  ],
      +  "type": "object"
      +}
    • changedInput schema / properties / duration / description
      Previous value: -"Duration as model-specific string enum, e.g. \"4s\", \"6s\", \"8s\". See GET /v1/models/:id/card."New value: +"Duration as model-specific string enum, e.g. \"4s\", \"6s\", \"8s\". See venice_model_details."
    • changedInput schema / properties / resolution / description
      Previous value: -"Output resolution, e.g. \"720p\", \"1080p\", \"4k\". Model-specific; see model card."New value: +"Output resolution, e.g. \"720p\", \"1080p\", \"4k\". Model-specific; see venice_model_details."
    • changedInput schema / required
      Previous value: -[
      -  "prompt",
      -  "model"
      -]New value: +[
      +  "prompt",
      +  "model",
      +  "duration"
      +]
  3. First observedv0.2.0

TDQS

A4.6/5.0
Behavior5/5

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

Beyond the annotations, it discloses the auth modes (x402 wallet or API key), the return shape, polling workflow, NSFW/model-policy behavior, model-specific duration constraints, and the defensive nature of Seedance consent flags. These details materially help an agent call and interpret the tool correctly.

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?

The description is front-loaded with purpose and next steps, and most sentences carry useful operational detail. It is somewhat dense and the long model list and repeated warnings add bulk, but the content is relevant for a 22-parameter generation tool.

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

Completeness5/5

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

For a complex tool with 22 parameters, nested objects, and no output schema, the description covers the essential workflow: queueing, return shape, polling, model selection, auth, and consent constraints. The schema carries detailed parameter semantics, so the description need not explain every field.

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

Parameters4/5

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

Schema coverage is high at 91%, so most parameter meaning is already documented. The description still adds value by giving concrete model id examples, warning that duration is a model-specific string enum, and clarifying that consent flags are defensive compatibility support rather than a policy bypass.

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?

The description opens with a specific verb and resource: "Queue a video generation." It distinguishes this from venice_video_status by explaining the returned queue_id and explicitly naming the status tool for polling.

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?

It gives clear routing guidance: poll with venice_video_status and inspect model-specific duration with venice_model_details. It also specifies when to set consent flags. It does not explicitly state when not to use this tool versus venice_video_quote or venice_video_complete, so it falls 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.