Skip to main content
Glama

Magic Hour

by ai.magichour.mcp
OverviewDiscussionsAdmin

Replace or animate a character using video and a reference image

character_replace_create_video

Use only with the user's own likeness or content they are authorized to use. Do not use for impersonation, deception, sexual content, or content involving minors.

MCP guidance:

  • Starts an async video job; returns id and credits_charged. For finished results, call wait_for_video_project with that id or poll GET /v1/video-projects/{id} until complete, error, or canceled. Results include downloads; wait helpers add exact_download_urls. Preserve signed URLs and query parameters; never append expiration metadata.

  • For *_file_path, use an existing Magic Hour path or the upload-URL endpoint's file_path after uploading raw bytes to its presigned URL. Direct public media URLs may work if stable and fetchable as raw bytes; hotlinked URLs can fail. Prefer the upload-URL endpoint when unsure.

Replace a video subject or transfer video motion onto a reference character image. Requires assets.video_file_path, assets.image_file_path and end_seconds. Model support controls duration, resolution and subject selection; see parameter schemas. Generation consumes credits; the queued video charge is an estimate, with the final charge available on the completed project.

Input Schema

NameRequiredDescriptionDefault
nameNoGive your video a custom name for easy identification.Character Replace - dateTime
modelNoModel to use. Defaults to `wan-animate`. * **`wan-animate`**: 480p, 720p. Supports `points` subject selection. * **`kling-3.0`**: 720p, 1080p. Clips of 3–10 seconds in `replace` mode or 3–30 seconds in `animate` mode. Picks the main person automatically, so `points` are rejected.wan-animate
styleNoOptional style controls for replace vs animate mode and subject selection.
assetsYesSource video and reference character image for the job.
resolutionNoOutput video resolution. Must be supported by `model`. Defaults to the lowest resolution available on your plan for that model.
end_secondsYesEnd time of your clip (seconds). Must be greater than start_seconds.
start_secondsNoStart time of your clip (seconds). Must be ≥ 0.

Output Schema

NameRequiredDescriptionDefault
idYesUnique ID of the video. Use it with the [Get video Project API](https://docs.magichour.ai/api-reference/video-projects/get-video-details) to fetch status and downloads.
credits_chargedYesThe amount of credits deducted from your account to generate the video. If the status is not 'complete', this value is an estimate and may be adjusted upon completion based on the actual FPS of the output video. If video generation fails, credits will be refunded, and this field will be updated to include the refund.

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed2 schema fields changed
    • changedInput schema / properties / end_seconds / example
      Previous value: -15New value: +5
    • changedInput schema / properties / resolution / example
      Previous value: -"720p"New value: +"480p"
  2. Changed5 schema fields changed
    • addedInput schema / properties / model
      Added value: +{
+  "default": "wan-animate",
+  "description": "Model to use. Defaults to `wan-animate`.\n\n* **`wan-animate`**: 480p, 720p. Supports `points` subject selection.\n* **`kling-3.0`**: 720p, 1080p. Clips of 3–10 seconds in `replace` mode or 3–30 seconds in `animate` mode. Picks the main person automatically, so `points` are rejected.",
+  "enum": [
+    "wan-animate",
+    "kling-3.0"
+  ],
+  "example": "wan-animate",
+  "type": "string"
+}
    • changedInput schema / properties / resolution / description
      Previous value: -"Output video resolution. Defaults to 480p, the lowest resolution available on your plan."New value: +"Output video resolution. Must be supported by `model`. Defaults to the lowest resolution available on your plan for that model."
    • changedInput schema / properties / resolution / enum
      Previous value: -[
-  "480p",
-  "720p"
-]New value: +[
+  "480p",
+  "720p",
+  "1080p"
+]
    • changedInput schema / properties / style / properties / points / description
      Previous value: -"On-frame markers for manual subject selection. Required when `selection_mode` is `point`. Ignored when `selection_mode` is `auto` or omitted."New value: +"On-frame markers for manual subject selection. Required when `selection_mode` is `point`. Ignored when `selection_mode` is `auto` or omitted. Rejected for models without subject selection (supported by `wan-animate`)."
    • changedInput schema / properties / style / properties / selection_mode / description
      Previous value: -"How to locate the subject in the source video. `auto` detects a person automatically. `point` uses your `points` to mark the subject. Defaults to `auto`."New value: +"How to locate the subject in the source video. `auto` detects a person automatically. `point` uses your `points` to mark the subject and is supported by `wan-animate`. Defaults to `auto`."
  3. Changed1 schema field changed
    • removedInput schema / properties / context
      Removed value: -{
-  "description": "Explain in 15-25 words, in third person, why this tool is called and how it supports the user's goal. For analytics only. You MUST describe only the abstract purpose of the tool call. NEVER include, repeat, paraphrase, or infer personal, sensitive, or identifying information from the user request or tool results, including names, emails, phone numbers, IPs, IDs, or credentials. You MUST generalize specific entities into roles such as \"a user\", \"the customer\", or \"an account\". Example: \"Retrieving a customer's recent orders to investigate a billing issue and help support determine the appropriate resolution.\"",
-  "type": "string"
-}
  4. Changed1 schema field changed
    • addedInput schema / properties / context
      Added value: +{
+  "description": "Explain in 15-25 words, in third person, why this tool is called and how it supports the user's goal. For analytics only. You MUST describe only the abstract purpose of the tool call. NEVER include, repeat, paraphrase, or infer personal, sensitive, or identifying information from the user request or tool results, including names, emails, phone numbers, IPs, IDs, or credentials. You MUST generalize specific entities into roles such as \"a user\", \"the customer\", or \"an account\". Example: \"Retrieving a customer's recent orders to investigate a billing issue and help support determine the appropriate resolution.\"",
+  "type": "string"
+}
  5. First observed

TDQS

A4.2/5.0
Behavior5/5

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

Goes well past the annotations: it discloses that the call starts an async job returning `id` and `credits_charged`, that credits are consumed with the queued charge being only an estimate while the final charge appears on the completed project, and that signed URLs must be preserved verbatim. These are exactly the cost and lifecycle traits an agent needs before invoking a mutable, credit-burning job.

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 nearly every sentence carries information; the safety preamble and the MCP guidance bullets both do work. Slight deduction because the actual purpose sentence lands third, after the policy and the async/credits guidance, which is a less front-loaded arrangement for a tool whose identity should be established first.

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?

With an output schema present and full param coverage, the description supplies the remaining gaps: async lifecycle, credits, and file-path sourcing. Minor residual gaps are durations/resolution compatibility, which it defers to the parameter schemas, and no explicit sibling routing, but nothing needed to call it correctly is absent.

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 earns above that by restating the three required inputs and explaining that model choice constrains duration, resolution and subject-selection support. It also adds practical meaning to the `*_file_path` params (existing Magic Hour path vs upload-URL `file_path` vs stable public URL) that the schema only gestures at.

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 pair ('Replace a video subject or transfer video motion onto a reference character image') and the required inputs, which distinguishes it from generic video tools like video_to_video_create_video. It never names a sibling explicitly (e.g. face_swap_create_video or body_swap_create_image), so an agent must infer the boundary from the replace/animate framing rather than from a direct contrast.

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 clear when-to-use context (subject replace vs motion transfer) plus operational guidance: prefer the upload-URL endpoint when unsure about file paths, and use wait_for_video_project rather than polling. It also gates use on authorization and prohibits impersonation/sexual/minors content. What's missing is a direct statement of when a different sibling (face_swap_create_video, body_swap_create_image) is the better choice.

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

All Magic Hour toolsInspect
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.

Resources