shapeless

Generate AI video

generate_footage

Destructive

Generate a short AI video clip from a prose prompt with one subject and one camera move, useful for b-roll or a single shot. Check existing footage first; each call spends credit.

Instructions

Generate one short AI video clip from a prompt: one continuous camera move over one subject, silent unless with_audio (more per second). The dearest asset there is; try find_footage first. The clip carries a SynthID watermark and needs an AI-generated label where the platform asks. Returns a media key. SPENDS CREDIT: about $0.18 of engine cost a default call, billed with the plan's margin; budgetUsd caps it (default 2, max 10) and chargedUsd says what it cost. Needs the write scope.

Input Schema

TableJSON Schema

Name	Required	Description
`name`	Yes	A 2-6 word filename for this clip, as you would describe it to a colleague - e.g. "rain on window neon". Becomes the file the user sees in their Assets.
`aspect`	No	Shape of the generated frame. 9:16 by default.
`prompt`	Yes	The shot as 1-3 sentences of flowing prose: subject, setting, light, and the ONE camera move. Say what the camera does, not what the story means.
`budgetUsd`	No	Optional: the most this call may spend, in dollars, default 2, max 10.
`resolution`	No	720p by default and right for b-roll under a cut. 1080p costs $0.05/s against $0.03/s - worth it only when the clip fills the frame and holds.
`with_audio`	No	Generate the clip's own sound too - $0.05/s at 720p against $0.03/s silent. Default false, because narration, music and sfx normally own the mix and a generated soundtrack fights them. Worth it when the clip sound IS the point: rain, a crowd, a machine, a diegetic moment that carries the shot. An unmuted clip is mixed into the master.
`duration_seconds`	No	4, 6, 8 (6 by default); anything else rounds up, and you are billed for the rounded length. You pay per second, and you will cut most of it.

Output Schema

TableJSON Schema

Name	Required	Description
`tool`	Yes
`media`	Yes
`output`	No	The tool's own result, shaped by the tool.
`chargedUsd`	Yes

Schema Changelog

Changes observed during successful MCP inspections.

First observedv0.2.4Oct 13, 2026

TDQS

A4.8/5.0

Behavior5/5

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

Goes well past the annotations: exact credit cost (~$0.18 default), budgetUsd cap behavior, chargedUsd reporting, required write scope, SynthID watermark, and platform AI-labeling obligation. The destructiveHint=true annotation is corroborated by the explicit spend warning rather than contradicted.

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?

Dense but front-loaded: the core action, the cheaper alternative, and the spend warning all come before the parameter caveats. Slightly run-on in the cost/watermark run of sentences, but each sentence carries distinct information.

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 paid, credit-spending, watermark-bearing generation tool, the description covers cost, cap mechanism, permissions, return handle ('Returns a media key'), and legal labeling. With an output schema present, it correctly avoids re-explaining the full return shape.

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 100%, so baseline is 3; the description still adds value by tying parameters to aggregate economics ('about $0.18 of engine cost a default call', budgetUsd default 2/max 10) and by framing duration as something 'you will cut most of'. It largely echoes the schema's per-second pricing, so it does not reach a 5.

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?

States a precise verb+resource+scope: 'Generate one short AI video clip from a prompt: one continuous camera move over one subject, silent unless with_audio'. It is clearly distinguishable from sibling find_footage (which it explicitly names) and from generate_image/edit_footage.

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

Usage Guidelines5/5

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

Routes explicitly: 'The dearest asset there is; try find_footage first', giving a when-not-this-tool condition. It also states when the with_audio variant is worth it ('when the clip sound IS the point: rain, a crowd, a machine').

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

Deploy Server