VideoGen

Generate video clip

generate_video_clip

Generate a video clip from a text prompt, source images, source videos, spokenDialogue, or reference audio. Typically takes 1–3 minutes (HIGH/MAX can be longer). Tell the user that wait up front and keep polling calmly.

Input Schema

TableJSON Schema

Name	Required	Description
`prompt`	No	Description of the video to generate.
`quality`	No	Video generation quality. Omit to use workspace settings.
`aspectRatio`	No	Output aspect ratio as width:height units (not pixels). Example: { width: 16, height: 9 }.
`audioFileIds`	No	Optional reference audio file ids used to lip-sync from a recording. Use spokenDialogue to have the model speak a line it generates itself.
`imageFileIds`	No	Reference images to guide the generated appearance. Use fileId values from upload_file, get_file, or list_files.
`videoFileIds`	No	Optional reference video file ids.
`generateAudio`	No	Whether the result must include generated audio.
`spokenDialogue`	No	Exact line the subject should speak as native lip-synced speech. The model synthesizes the voice from this text.
`durationSeconds`	No	Requested clip playback duration in seconds. Omit or pass null for automatic timing.
`startFrameFileId`	No	Optional opening-frame still file id. Used as the first frame of the clip.
`voiceDescription`	No	Natural-language description of the voice that speaks spokenDialogue. Used when spokenDialogue is set.
`suppressBackgroundMusic`	No	When true, the clip will not include a musical soundtrack. Spoken dialogue and environmental sound are still allowed. Use this when background music will be added separately.

Output Schema

TableJSON Schema

Name	Required	Description
`error`	No	Error details when status is failed; otherwise null.
`status`	No	Job status: pending, running, succeeded, failed, or cancelled.
`results`	No	Generated results with download URLs when succeeded.
`toolType`	No	Tool name (e.g. GENERATE_IMAGE).
`attemptIndex`	No	Current or latest attempt index.
`toolExecutionId`	Yes	Tool execution id (vg_tool_...).
`progressPercentage`	No	Completion progress 0-100 (present after polling).

Schema Changelog

Changes observed during successful MCP inspections.

Changed2 schema fields changedOct 16, 2026

changedInput schema / properties / durationSeconds / description

Previous value: -"Optional clip length in whole seconds (1 to 30). Omit or pass null for Auto."New value: +"Requested clip playback duration in seconds. Omit or pass null for automatic timing."

changedInput schema / properties / imageFileIds / description

Previous value: -"Optional reference image file ids."New value: +"Reference images to guide the generated appearance. Use fileId values from upload_file, get_file, or list_files."

First observedOct 9, 2026

TDQS

A3.7/5.0

Behavior4/5

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

Annotations only cover the mutation profile (readOnlyHint=false, idempotentHint=false, destructiveHint=false). The description adds genuinely non-structured context: this is a long-running async operation taking 1–3 minutes (longer for HIGH/MAX) that must be polled, which is exactly the kind of trait an agent needs and annotations do not carry.

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?

Three sentences, front-loaded with the verb and resource. The latency/polling sentence is actionable and free of filler; slight redundancy in listing every input modality, but overall tight.

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?

For a 12-parameter async generation tool with an output schema and full schema coverage, the description supplies the missing piece – expected latency and the need to poll – without needing to explain returns. It stops short of naming the actual polling mechanism (e.g. get_tool_execution / get_async_tasks_guidance), which is the one residual gap.

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

Parameters3/5

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

Schema description coverage is 100%, so all 12 parameters are already documented in the schema. The description names a few modalities that map to parameters (spokenDialogue, source images/videos, reference audio) but adds no syntax, ordering, or dependency detail beyond what the schema provides – baseline 3.

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

Purpose4/5

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

States a specific verb+resource ('Generate a video clip') and enumerates the input modalities (text prompt, source images, source videos, spokenDialogue, reference audio). It is clearly a generation tool, but it never distinguishes itself from the many video-generation siblings (prompt_to_video_clip, script_to_video, voiceover_to_video), so an agent still has to guess which generator to pick.

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

Usage Guidelines3/5

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

The only guidance is operational ('Tell the user that wait up front and keep polling calmly'), which addresses handling rather than selection. It never says when to use this tool instead of prompt_to_video_clip or the other *_to_video siblings, leaving the choice implicit.

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

All VideoGen tools Inspect

Try in Browser

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.