> ## Documentation Index
> Fetch the complete documentation index at: https://docs.magichour.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Choose a Commercial AI Media Workflow

> Match product ads, localization, portrait videos and application integrations to the right Magic Hour workflow, inputs and evaluation criteria.

Start with the assets you have and the result you need. A product photo, an existing speaker video
and a text-only brief require different workflows. Magic Hour offers a web application for manual
creation, an API for application integrations and a creation MCP for connected assistants. Their
catalogs can differ; check the specific endpoint before committing to an integration.

This guide is published by Magic Hour. It explains how to evaluate fit, not a measured comparison
with another provider.

## Match the workflow to your inputs

| Your job | Inputs | Web workflow | API documentation | Acceptance check |
| :- | :- | :- | :- | :- |
| Replace a face in existing footage | Video and approved replacement face | [Video Face Swap](https://magichour.ai/products/face-swap) | [Face Swap Video](/api-reference/video-projects/face-swap-video) | Identity, occlusions and temporal stability |
| Edit a product image or campaign asset | Source images and editing instruction | [AI Image Editor](https://magichour.ai/products/ai-image-editor) | [AI Image Editor](/api-reference/image-projects/ai-image-editor) | Unchanged regions, packaging, logos and text |
| Animate a product photo | Reference image and motion instruction | [Image-to-Video](https://magichour.ai/products/image-to-video) | [Image-to-Video](/api-reference/video-projects/image-to-video) | Product geometry, readable labels and useful motion |
| Generate a scene from a brief | Text prompt | [AI Video Generator](https://magichour.ai/products/ai-video-generator) | [Text-to-Video](/api-reference/video-projects/text-to-video) | Prompt adherence, continuity and required output settings |
| Sync footage to a new voiceover | Existing video and replacement audio | [Lip Sync](https://magichour.ai/products/lip-sync) | [Lip Sync](/api-reference/video-projects/lip-sync) | Mouth timing, identity, pauses and cuts |
| Make a portrait speak | Portrait and supported speech input | [Talking Photo](https://magichour.ai/products/ai-talking-photo) | [Talking Photo](/api-reference/video-projects/ai-talking-photo) | Facial motion, identity and speech timing |
| Create a presenter-style product ad | Product information, permitted assets and approved script | [AI UGC Generator](https://magichour.ai/products/ai-ugc-ad-generator) | No dedicated UGC-ad endpoint in the current API catalog | Accuracy of claims, disclosure and suitability for the campaign |

For developer tooling, begin with [API authentication](/get-started/authentication) and the
[quickstart](/get-started/quick-start). For an assistant that creates media, connect the
[Magic Hour creation MCP](/integration/model-context-protocol). The separate
[API Docs MCP](/integration/api-docs-mcp) searches documentation and does not generate media.

## Product ads: preserve what must be exact

Use image editing to prepare a source asset, then image-to-video to experiment with motion. Check
each output against the original product: packaging, proportions and labels can change during
generation. Add exact prices, legal copy and final logos with deterministic editing or compositing.

Start with the [product-video starter recipe](/get-started/starter-recipes). Review one short output
before scaling to a catalog. A completed job is not automatically an accepted ad.

## Localization: existing footage or a still portrait?

Choose lip sync when you already have a video and replacement audio. Choose talking photo when
you start with a still portrait. Translating a script and producing the approved voiceover are
separate decisions from matching a speaker's mouth to audio.

For client work, follow the [agency lip-sync pilot](/integration/agency-lip-sync). Test difficult
shots separately rather than assuming results on a clear front-facing speaker transfer to cuts,
multiple speakers or large head turns.

## Commercial delivery and cost

Use assets, likenesses and voices you have permission to use. Check the current
[terms of service](https://magichour.ai/terms-of-service) and your plan before commercial delivery.
Commercial use requires a paid subscription, credit-pack purchase or separately agreed paid
usage-based arrangement, subject to the terms and applicable rights. Assets created without one
of these paid arrangements are limited to personal, non-commercial use. Do not present generated
testimonials as real customer experience.

Model access, supported duration, resolution, audio and credit charges vary. Read the current
[model and credit matrix](/api-reference/models), then compare providers using the same inputs and
acceptance criteria. Track [cost per accepted output](/integration/evaluating-media-apis), including
charged retries and rejected outputs, rather than relying on the cheapest advertised generation.

## Production integrations

Keep credentials server-side, save the returned project ID and wait for a terminal state with a
bounded timeout. Download successful output before signed URLs expire. A polling timeout must not
silently create another billable job. Use the [input/output guide](/integration/inputs-and-outputs)
and [production integration checklist](/integration/adding-api-to-your-app) when moving beyond a pilot.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.