google-flow-mcp

Overview Schema Source Related Servers Discussions Admin

flow_extend_video

Continue or edit a video clip via prompts: extend scenes by 3–10 seconds per turn up to 30 seconds, or change visual details such as sky color.

Instructions

Continue or edit a video with Gemini Omni Flash.

Best: pass previous_interaction_id from an earlier video result (keeps full context, supports new dialogue). Otherwise pass video_path to upload a clip of 15 s or less. Extension adds 3-10 s per turn, up to 30 s total. Prompts like "Make the sky purple" edit the clip instead of extending it.

Args: prompt: How the scene continues, or what to change. video_path: Local MP4 to extend (uploaded through the Files API). previous_interaction_id: interaction_id printed by an earlier video tool. duration_seconds: Length of the new part, 3-10. resolution: '360p', '720p', '1080p' or '4k'. reference_image_paths: Images of new characters/objects to bring in. confirm_cost: Set true only after the user accepted the cost estimate. output_dir: Override the output folder. aspect_ratio: '16:9' or '9:16'. Usually not needed: a local clip's shape is detected. Pass it if a vertical clip comes back landscape.

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>"
    },
    "name": "flow_extend_video"
  }
}

Input Schema

Name	Required	Default
`prompt`	Yes
`output_dir`	No
`resolution`	No	720p
`video_path`	No
`aspect_ratio`	No
`confirm_cost`	No
`duration_seconds`	No
`reference_image_paths`	No
`previous_interaction_id`	No

Schema Changelog

Changes observed during successful MCP inspections.

First observedv0.3.3Nov 13, 2026

TDQS

A4.3/5.0

Behavior4/5

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

No annotations are supplied, so the description carries the full burden, and it does disclose real behavior: uploads go through the Files API, extension is capped per-turn and in total, and confirm_cost must only be set after a user accepts a cost estimate — a meaningful guardrail. It omits whether the call is long-running/async and what the response contains, which matters for a generation tool.

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

Conciseness5/5

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

Front-loads the decision rule ('Best: ... Otherwise ...') before the parameter list, so an agent can act on the first two lines. Every sentence carries information — mode selection, limits, cost gate, or a parameter constraint — with no filler.

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 9-parameter, no-annotation, no-output-schema generation tool, this is nearly complete: input sourcing, duration/resolution constraints, cost confirmation, and output location are all covered. The main gap is the return shape (the interaction_id is referenced as an input from 'an earlier video tool' but never described as this tool's output) and whether the call blocks.

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?

Schema description coverage is 0% across 9 parameters, and the Args block compensates for every one: allowed resolution values (including '4k'), the 3-10 duration range, reference_image_paths being characters/objects to bring in, output_dir as an override, and the confirm_cost gating rule. The aspect_ratio note even explains the common failure it fixes ('if a vertical clip comes back landscape').

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?

Opens with a specific verb+resource: 'Continue or edit a video', and immediately disambiguates the two modes ('Prompts like "Make the sky purple" edit the clip instead of extending it'), which is the key mental model. It does not explicitly contrast itself with the sibling flow_edit_video, so an agent must infer the boundary rather than read it.

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 an explicit priority order for input modes: 'Best: pass previous_interaction_id ... Otherwise pass video_path', including the 15 s constraint on uploads, plus concrete limits (3-10 s per turn, 30 s total) and a conditional rule for aspect_ratio. No named alternatives among siblings (flow_edit_video, flow_generate_video) and no explicit when-not-to-use statement.

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