Magic Hour

AI Talking Photo

ai_talking_photo_create_talking_photo

Create a talking photo video from an image and an audio file.

MCP guidance:

This starts an async video generation job and returns id plus credits_charged immediately. If the user wants the finished result, call the wait_for_video_project helper with the returned id, or poll the matching GET /v1/video-projects/{id} endpoint until status is complete, error, or canceled. Completed projects include downloads with direct URLs. The custom wait helper also returns exact_download_urls separately from expiration metadata.
For *_file_path values, prefer an existing Magic Hour file path or a file_path returned by the upload-URL endpoint after the file bytes are uploaded. Direct public media URLs may work when they are stable, fetchable, and return raw file bytes, but hotlinked URLs can fail; when in doubt, use the presigned upload flow first and pass the returned file_path.

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.

Input Schema

TableJSON Schema

Name	Required	Description	Default
`name`	No	Give your video a custom name for easy identification.	Talking Photo - dateTime
`style`	No	Attributes used to dictate the style of the output
`assets`	Yes	Provide the assets for creating a talking photo
`end_seconds`	Yes	The end time of the input audio in seconds. Maximum clip length depends on style.generation_mode: realistic 300s, prompted 45s.
`start_seconds`	Yes	The start time of the input audio in seconds. Maximum clip length depends on style.generation_mode: realistic 300s, prompted 45s.
`max_resolution`	No	Constrains the larger dimension (height or width) of the output video. Allows you to set a lower resolution than your plan's maximum if desired. The value is capped by your plan's max resolution.

Output Schema

TableJSON Schema

Name	Required	Description	Default
`id`	Yes	Unique 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_charged`	Yes	The 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.

Changed2 schema fields changedOct 7, 2026

changedInput schema / properties / name / description

Previous value: -"Give your image a custom name for easy identification."New value: +"Give your video a custom name for easy identification."

changedInput schema / properties / name / example

Previous value: -"My Talking Photo image"New value: +"My Talking Photo video"

Changed1 schema field changedSep 24, 2026

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"
-}

Changed1 schema field changedSep 11, 2026

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"
+}

Changed2 schema fields changedSep 2, 2026

changedInput schema / properties / end_seconds / description

Previous value: -"The end time of the input audio in seconds. Maximum clip length depends on style.generation_mode: realistic 180s, prompted 45s."New value: +"The end time of the input audio in seconds. Maximum clip length depends on style.generation_mode: realistic 300s, prompted 45s."

changedInput schema / properties / start_seconds / description

Previous value: -"The start time of the input audio in seconds. Maximum clip length depends on style.generation_mode: realistic 180s, prompted 45s."New value: +"The start time of the input audio in seconds. Maximum clip length depends on style.generation_mode: realistic 300s, prompted 45s."

First observedSep 1, 2026

TDQS

A4/5.0

Behavior5/5

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

Goes well beyond the annotations (which only say it is a non-read-only, open-world write): it discloses that it starts an async job returning `id` and `credits_charged` immediately, documents terminal statuses, notes that completed projects carry `downloads`/`exact_download_urls`, and adds an authorization/acceptable-use restriction. This is the kind of cost and lifecycle context an agent needs and cannot get from structured fields.

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?

Purpose is front-loaded in one sentence, followed by clearly delimited guidance bullets. Slightly long, and the return-metadata detail partly duplicates what an output schema could cover, but no sentence is wasted.

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 an async, nested-object, 6-parameter generation tool with an output schema, the description covers the full call lifecycle: trigger, wait/poll, and completion. It does not cover failure modes or credit/pricing limits beyond the immediate charge, leaving minor gaps.

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?

With 100% schema coverage the baseline is 3, but the description adds real value on `*_file_path`: it explains the preference for Magic Hour file paths or upload-URL `file_path` values and warns that hotlinked public URLs may fail. That is guidance beyond the schema's field descriptions.

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 and resource: 'Create a talking photo video from an image and an audio file.' Inputs are explicit, so an agent can distinguish it from general image/video generators. It does not explicitly name the closest sibling (lip_sync_create_video), so it falls short of full sibling differentiation.

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?

Provides strong workflow guidance (async job, how to wait/poll) and a clear authorization constraint, but never states when to pick this tool over related siblings like lip_sync_create_video or image_to_video_create_video. Usage is implied by the input shape rather than spelled out as a selection rule.

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

All Magic Hour 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.