# Magic Hour > Create and edit images, videos, and audio through Magic Hour's hosted Streamable HTTP MCP server. - name: io.github.magichourhq/magic-hour - version: 0.1.0 - connection class: R0 (autonomous) - trust: 91/100 - quality: 70/100 (needs work) - owner: not claimed - descriptor: https://api.protogrid.dev/v1/servers/io.github.magichourhq%2Fmagic-hour - tools: https://api.protogrid.dev/v1/servers/io.github.magichourhq%2Fmagic-hour/tools **An agent can connect right now, no human step.** Remote endpoint, no authentication, reachable on the last probe. ## Connection (mcpServers) ```json { "mcpServers": { "magic-hour": { "type": "http", "url": "https://mcp.magichour.ai/" } } } ``` ## Trust - hygiene: 100 - liveness: 100 - freshness: 85 - provenance: 80 - drivers: +repository +namespace-matches-repo +reachable +uptime-100% ~updated-51d-ago - Derived from observable signals (official registry feed and our own credential-free probes); no code audit performed. ## Quality 70/100 (needs work) ### protocol: 75 (weight 25) - warn `protocol.modern`: newest supported version is 2025-11-25; 2026-07-28 not supported, older versions are deprecated until 2027-07-28 - n/a `protocol.stateless`: only applies to 2026-07-28 servers - pass `protocol.transport`: streamable HTTP - n/a `protocol.list_ttl`: only applies to 2026-07-28 servers ### authorization: n/a (weight 15) - n/a `auth.prm`, `auth.as_metadata`, `auth.cimd`: no remote uses OAuth - n/a `auth.secret_in_url`: no templated URL ### tool hygiene: 64 (weight 30) - pass `tools.descriptions`: every tool has a description - warn `tools.description_length`: 10 tools have descriptions over 1,024 characters, which crowds the context - pass `tools.schemas`: every tool has an object input schema - pass `tools.annotations`: 46 of 46 tools declare readOnlyHint or destructiveHint - fail `tools.directory_hints`: no tool declares all four of readOnlyHint, destructiveHint, idempotentHint and openWorldHint; missing: wait_for_audio_project (idempotentHint); fetch_audio_download (idempotentHint); fetch_video_download (idempotentHint); … - fail `tools.token_cost`: about 32,355 tokens to load every tool, a large share of an agent's context - pass `tools.api_dump`: tools are not a one-to-one API dump ### stability: 75 (weight 15) - warn `stability.changes`: 279 tool changes in 30 days - pass `stability.rug_pull`: no tool changed its meaning under the same name ### dependencies: n/a (weight 15) - n/a `deps.known_vulns`, `deps.mcp_sdk_version`, `deps.resolvable`: no npm or PyPI package - tools: 46, about 32,355 tokens to load them all - tool set hash: 4aaa41ee2bfdc7cfef84f2ee0c452ce5 - tools last changed: 2026-10-18T05:02:17.905Z - badge: [![protogrid quality](https://protogrid.dev/badge/io.github.magichourhq/magic-hour.svg)](https://protogrid.dev/servers/io.github.magichourhq/magic-hour) - Checks run on what our credential-free, read-only probes observe; tools are never called and no code audit is performed. ### Tool changes - 2026-10-18 `workflow_guide_retrieve` changed (description) - before: Read Magic Hour workflow guidance for face swap, image editing, image-to-video, video generation, lip sync, UGC ads, talking photos or API integration. Returns product and documentation URLs, exact tool names and review criteria; does not generate media, spend credits or check account access. - after: Read Magic Hour workflow guidance for face swap, image editing, image-to-video, video generation, lip sync, UGC ads, talking photos or API integration. Returns product and documentation URLs, exact tool names, ordered steps, example requests, follow-up edits and review criteria; does not generate media, spend credits or check account access. - 2026-10-17 `ai_image_editor_create_image` changed (description) - before: MCP guidance: - Starts an async image job; returns `id` and `credits_charged`. For finished results, call `wait_for_image_project` with that id or poll `GET /v1/image-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. Edit images with AI. - after: MCP guidance: - Requires one or more source images in `assets.image_file_paths`. For prompt-only creation, use `ai_image_generator_create_image` instead. - Starts an async image job; returns `id` and `credits_charged`. For finished results, call `wait_for_image_project` with that id or poll `GET /v1/image-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. Edit images with AI. - 2026-10-17 `fetch_video_download` changed (description, inputSchema) - before: Fetch a video `downloads[n].url` from a completed video project and return it as an embedded MCP binary resource for compatible clients. Pass the exact full signed URL from `downloads[n].url` without trimming query parameters; `expires_at` is separate metadata, not part of the URL. - after: Fetch a video `downloads[n].url` from a completed video project and return it as an embedded MCP binary resource for compatible clients. Pass the exact full signed URL from `downloads[n].url` without trimming query parameters; `expires_at` is separate metadata, not part of the URL. Inline media is capped at 8 MiB; use the signed URL directly for larger downloads. - 2026-10-17 `fetch_audio_download` changed (description, inputSchema) - before: Fetch a audio `downloads[n].url` from a completed audio project and return it as inline MCP audio content for compatible clients. Pass the exact full signed URL from `downloads[n].url` without trimming query parameters; `expires_at` is separate metadata, not part of the URL. - after: Fetch a audio `downloads[n].url` from a completed audio project and return it as inline MCP audio content for compatible clients. Pass the exact full signed URL from `downloads[n].url` without trimming query parameters; `expires_at` is separate metadata, not part of the URL. Inline media is capped at 8 MiB; use the signed URL directly for larger downloads. - 2026-10-17 `fetch_image_download` changed (description, inputSchema) - before: Fetch a image `downloads[n].url` from a completed image project and return it as inline MCP image content for compatible clients. Pass the exact full signed URL from `downloads[n].url` without trimming query parameters; `expires_at` is separate metadata, not part of the URL. - after: Fetch a image `downloads[n].url` from a completed image project and return it as inline MCP image content for compatible clients. Pass the exact full signed URL from `downloads[n].url` without trimming query parameters; `expires_at` is separate metadata, not part of the URL. Inline media is capped at 8 MiB; use the signed URL directly for larger downloads. - 2026-10-17 `wait_for_audio_project` changed (description, inputSchema) - before: Poll an audio project until it completes, errors, is canceled, or times out. Returns the final project JSON and, when complete, attempts to inline audio downloads for Inspector or compatible clients. Returns sanitized download fields. Use `exact_download_urls[n]` or `downloads[n].url` exactly as returned; do not shorten it, remove query parameters, or append expiration metadata. - after: Poll an audio project until it completes, errors, is canceled, or times out. After a timeout, resume with the same id; do not create another generation. Returns the final project JSON and, when complete, attempts to inline audio downloads for Inspector or compatible clients. Returns sanitized download fields. Use `exact_download_urls[n]` or `downloads[n].url` exactly as returned; do not shorten it, remove query parameters, or append expiration metadata. Inline previews share an 8 MiB limit; all download URLs remain available. - 2026-10-17 `wait_for_image_project` changed (description, inputSchema) - before: Poll an image project until it completes, errors, is canceled, or times out. Returns the final project JSON and, when complete, attempts to inline image downloads for Inspector or compatible clients. Returns sanitized download fields. Use `exact_download_urls[n]` or `downloads[n].url` exactly as returned; do not shorten it, remove query parameters, or append expiration metadata. - after: Poll an image project until it completes, errors, is canceled, or times out. After a timeout, resume with the same id; do not create another generation. Returns the final project JSON and, when complete, attempts to inline image downloads for Inspector or compatible clients. Returns sanitized download fields. Use `exact_download_urls[n]` or `downloads[n].url` exactly as returned; do not shorten it, remove query parameters, or append expiration metadata. Inline previews share an 8 MiB limit; all download URLs remain available. - 2026-10-17 `wait_for_video_project` changed (description) - before: Poll a video project until it completes, errors, is canceled, or times out. Returns sanitized download fields. Use `exact_download_urls[n]` or `downloads[n].url` exactly as returned; do not shorten it, remove query parameters, or append expiration metadata. - after: Poll a video project until it completes, errors, is canceled, or times out. After a timeout, resume with the same id; do not create another generation. Returns sanitized download fields. Use `exact_download_urls[n]` or `downloads[n].url` exactly as returned; do not shorten it, remove query parameters, or append expiration metadata. Inline previews share an 8 MiB limit; all download URLs remain available. - 2026-10-17 `ai_ugc_ad_create_video` added - 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. Create a spoken UGC video from a prepared image and script. Returns a project ID; poll GET /v1/video-projects/{id} for status and downloads. To compose an image, generate the image separately with the Image Editor API and supply its download URL. - 2026-10-15 `audio_projects_retrieve_details` changed (description) - before: Check the progress of a audio project. The `downloads` field is populated after a successful render. **Statuses** - `queued` — waiting to start - `rendering` — in progress - `complete` — ready; see `downloads` - `error` — a failure occurred (see `error`) - `canceled` — user canceled - `draft` — not used MCP guidance: - Use this after a create tool to poll job status. When status is `complete`, surface the `downloads` URLs to the user; if status is `error`, surface the error message. - Each `downloads[n].url` is already the full signed download URL. Use it exactly as returned. Do not shorten it, strip query parameters, or append `expires_at` onto the URL string. - after: MCP guidance: - Use this after a create tool to poll job status. When status is `complete`, surface the `downloads` URLs to the user; if status is `error`, surface the error message. - Each `downloads[n].url` is already the full signed download URL. Use it exactly as returned. Do not shorten it, strip query parameters, or append `expires_at` onto the URL string. Check the progress of a audio project. The `downloads` field is populated after a successful render. **Statuses** - `queued` — waiting to start - `rendering` — in progress - `complete` — ready; see `downloads` - `error` — a failure occurred (see `error`) - `canceled` — user canceled - `draft` — not used Full history: https://api.protogrid.dev/v1/servers/io.github.magichourhq%2Fmagic-hour/changes ## Remotes - https://mcp.magichour.ai/ (streamable-http, auth none, reachable true, uptime30d 1) ## Packages _none_ ## Tools (46) - `account_retrieve`: Get the current credit balance and subscription details of the account that owns the API key. - `ai_clothes_changer_create_image`: 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 image job; returns `id` and `credits_charged`. For finished results, call `wait_for_image_project` with that id or poll `GET /v1/image-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. Change outfits in photos in seconds with just a photo reference. Each photo costs 25 credits. - `ai_face_editor_edit_image`: MCP guidance: - Starts an async image job; returns `id` and `credits_charged`. For finished results, call `wait_for_image_project` with that id or poll `GET /v1/image-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. Edit facial features of an image using AI. Each edit costs 1 frame. The height/width of the output image depends on your subscription. Please refer to our [pricing](https://magichour.ai/pricing) page for more details - `ai_gif_generator_create_image`: MCP guidance: - Starts an async image job; returns `id` and `credits_charged`. For finished results, call `wait_for_image_project` with that id or poll `GET /v1/image-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. Create an AI GIF. Each GIF costs 50 credits. - `ai_headshot_generator_create_image`: MCP guidance: - Starts an async image job; returns `id` and `credits_charged`. For finished results, call `wait_for_image_project` with that id or poll `GET /v1/image-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. Create an AI headshot. Each headshot costs 50 credits. - `ai_image_editor_create_image`: MCP guidance: - Requires one or more source images in `assets.image_file_paths`. For prompt-only creation, use `ai_image_generator_create_image` instead. - Starts an async image job; returns `id` and `credits_charged`. For finished results, call `wait_for_image_project` with that id or poll `GET /v1/image-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. Edit images with AI. - `ai_image_generator_create_image`: MCP guidance: - Starts an async image job; returns `id` and `credits_charged`. For finished results, call `wait_for_image_project` with that id or poll `GET /v1/image-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. Create an AI image with advanced model selection and quality controls. - `ai_image_upscaler_create_image`: MCP guidance: - Starts an async image job; returns `id` and `credits_charged`. For finished results, call `wait_for_image_project` with that id or poll `GET /v1/image-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. Upscale your image using AI. Each 2x upscale costs 50 credits for balanced/creative modes, and 25 credits for preserve. 4x upscale costs 200 and 100 credits respectively. - `ai_meme_generator_create_image`: MCP guidance: - Starts an async image job; returns `id` and `credits_charged`. For finished results, call `wait_for_image_project` with that id or poll `GET /v1/image-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. Create an AI generated meme. Each meme costs 10 credits. - `ai_qr_code_generator_create_image`: MCP guidance: - Starts an async image job; returns `id` and `credits_charged`. For finished results, call `wait_for_image_project` with that id or poll `GET /v1/image-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. Create an AI QR code. Each QR code costs 0 credits. - `ai_talking_photo_create_talking_photo`: 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. Create a talking photo video from an image and an audio file. - `ai_ugc_ad_create_video`: 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. Create a spoken UGC video from a prepared image and script. Returns a project ID; poll GET /v1/video-projects/{id} for status and downloads. To compose an image, generate the image separately with the Image Editor API and supply its download URL. - `ai_video_editor_create_video`: 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. **What this API does** Create the same Video Editor you can make in the browser, but programmatically, so you can automate it, run it at scale, or connect it to your own app or workflow. **Good for** - Automation and batch processing - Adding video editor into apps, pipelines, or tools **How it works (3 steps)** 1) Upload your inputs (video, image, or audio) with [Generate Upload URLs](https://docs.magichour.ai/api-reference/files/generate-asset-upload-urls) and copy the `file_path`. 2) Send a request to create a video editor job with the basic fields. 3) Check the job status until it's `complete`, then download the result from `downloads`. **Key options** - Inputs: usually a file, sometimes a YouTube link, depending on project type - Resolution: free users are limited to 576px; higher plans unlock HD and larger sizes - Extra fields: e.g. `face_swap_mode`, `start_seconds`/`end_seconds`, or a text prompt **Cost** Credits are only charged for the frames that actually render. You'll see an estimate when the job is queued, and the final total after it's done. For detailed examples, see the [product page](https://magichour.ai/products/ai-video-editor). - `ai_video_translator_create_video`: 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. **What this API does** Create the same Video Translator you can make in the browser, but programmatically, so you can automate it, run it at scale, or connect it to your own app or workflow. **Good for** - Automation and batch processing - Adding video translator into apps, pipelines, or tools **How it works (3 steps)** 1) Upload your inputs (video, image, or audio) with [Generate Upload URLs](https://docs.magichour.ai/api-reference/files/generate-asset-upload-urls) and copy the `file_path`. 2) Send a request to create a video translator job with the basic fields. 3) Check the job status until it's `complete`, then download the result from `downloads`. **Key options** - Inputs: usually a file, sometimes a YouTube link, depending on project type - Resolution: free users are limited to 576px; higher plans unlock HD and larger sizes - Extra fields: e.g. `face_swap_mode`, `start_seconds`/`end_seconds`, or a text prompt **Cost** Credits are only charged for the frames that actually render. You'll see an estimate when the job is queued, and the final total after it's done. For detailed examples, see the [product page](https://magichour.ai/products/ai-video-translator). - `animation_create_video`: 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. Create a Animation video. The estimated frame cost is calculated based on the `fps` and `end_seconds` input. - `audio_projects_delete`: Permanently delete the rendered audio file(s). This action is not reversible, please be sure before deleting. - `audio_projects_retrieve_details`: MCP guidance: - Use this after a create tool to poll job status. When status is `complete`, surface the `downloads` URLs to the user; if status is `error`, surface the error message. - Each `downloads[n].url` is already the full signed download URL. Use it exactly as returned. Do not shorten it, strip query parameters, or append `expires_at` onto the URL string. Check the progress of a audio project. The `downloads` field is populated after a successful render. **Statuses** - `queued` — waiting to start - `rendering` — in progress - `complete` — ready; see `downloads` - `error` — a failure occurred (see `error`) - `canceled` — user canceled - `draft` — not used - `audio_to_video_create_video`: 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. **What this API does** Create the same Audio To Video you can make in the browser, but programmatically, so you can automate it, run it at scale, or connect it to your own app or workflow. **Good for** - Automation and batch processing - Adding audio to video into apps, pipelines, or tools **How it works (3 steps)** 1) Upload your inputs (video, image, or audio) with [Generate Upload URLs](https://docs.magichour.ai/api-reference/files/generate-asset-upload-urls) and copy the `file_path`. 2) Send a request to create a audio to video job with the basic fields. 3) Check the job status until it's `complete`, then download the result from `downloads`. **Key options** - Inputs: usually a file, sometimes a YouTube link, depending on project type - Resolution: free users are limited to 576px; higher plans unlock HD and larger sizes - Extra fields: e.g. `face_swap_mode`, `start_seconds`/`end_seconds`, or a text prompt **Cost** Credits are only charged for the frames that actually render. You'll see an estimate when the job is queued, and the final total after it's done. For detailed examples, see the [product page](https://magichour.ai/products/audio-to-video). - `auto_subtitle_generator_create_video`: 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. Automatically generate subtitles for your video in multiple languages. - `body_swap_create_image`: 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 image job; returns `id` and `credits_charged`. For finished results, call `wait_for_image_project` with that id or poll `GET /v1/image-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. Swap a person into a scene image using Nano Banana 2 Lite (640px/1k) or Nano Banana 2 (2k/4k). Credits depend on `resolution` (from 50 credits at 640px upward). - `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. **What this API does** Create the same Character Replace you can make in the browser, but programmatically, so you can automate it, run it at scale, or connect it to your own app or workflow. **Good for** - Automation and batch processing - Adding character replace into apps, pipelines, or tools **How it works (3 steps)** 1) Upload your inputs (video, image, or audio) with [Generate Upload URLs](https://docs.magichour.ai/api-reference/files/generate-asset-upload-urls) and copy the `file_path`. 2) Send a request to create a character replace job with the basic fields. 3) Check the job status until it's `complete`, then download the result from `downloads`. **Key options** - Inputs: usually a file, sometimes a YouTube link, depending on project type - Resolution: free users are limited to 576px; higher plans unlock HD and larger sizes - Extra fields: e.g. `face_swap_mode`, `start_seconds`/`end_seconds`, or a text prompt **Cost** Credits are only charged for the frames that actually render. You'll see an estimate when the job is queued, and the final total after it's done. For detailed examples, see the [product page](https://magichour.ai/products/character-replace). - `face_detection_detect_faces`: MCP guidance: - This starts an async face-detection task and returns an `id`. Use the face-detection details endpoint with that id to retrieve detected faces before doing individual face swaps. - 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. Detect faces in an image or video. Use this API to get the list of faces detected in the image or video to use in the [face swap photo](https://docs.magichour.ai/api-reference/image-projects/face-swap-photo) or [face swap video](https://docs.magichour.ai/api-reference/video-projects/face-swap-video) API calls for multi-face swaps. Note: Face detection is free to use for the near future. Pricing may change in the future. - `face_detection_retrieve_details`: Get the details of a face detection task. Use this API to get the list of faces detected in the image or video to use in the [face swap photo](https://docs.magichour.ai/api-reference/image-projects/face-swap-photo) or [face swap video](https://docs.magichour.ai/api-reference/video-projects/face-swap-video) API calls for multi-face swaps. - `face_swap_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. **What this API does** Create the same Face Swap you can make in the browser, but programmatically, so you can automate it, run it at scale, or connect it to your own app or workflow. **Good for** - Automation and batch processing - Adding face swap into apps, pipelines, or tools **How it works (3 steps)** 1) Upload your inputs (video, image, or audio) with [Generate Upload URLs](https://docs.magichour.ai/api-reference/files/generate-asset-upload-urls) and copy the `file_path`. 2) Send a request to create a face swap job with the basic fields. 3) Check the job status until it's `complete`, then download the result from `downloads`. **Key options** - Inputs: usually a file, sometimes a YouTube link, depending on project type - Resolution: free users are limited to 576px; higher plans unlock HD and larger sizes - Extra fields: e.g. `face_swap_mode`, `start_seconds`/`end_seconds`, or a text prompt **Cost** Credits are only charged for the frames that actually render. You'll see an estimate when the job is queued, and the final total after it's done. For detailed examples, see the [product page](https://magichour.ai/products/face-swap). - `face_swap_photo_create_image`: 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 image job; returns `id` and `credits_charged`. For finished results, call `wait_for_image_project` with that id or poll `GET /v1/image-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. Create a face swap photo. Each photo costs 10 credits. The height/width of the output image depends on your subscription. Please refer to our [pricing](https://magichour.ai/pricing) page for more details - `fetch_audio_download`: Fetch a audio `downloads[n].url` from a completed audio project and return it as inline MCP audio content for compatible clients. Pass the exact full signed URL from `downloads[n].url` without trimming query parameters; `expires_at` is separate metadata, not part of the URL. Inline media is capped at 8 MiB; use the signed URL directly for larger downloads. - `fetch_image_download`: Fetch a image `downloads[n].url` from a completed image project and return it as inline MCP image content for compatible clients. Pass the exact full signed URL from `downloads[n].url` without trimming query parameters; `expires_at` is separate metadata, not part of the URL. Inline media is capped at 8 MiB; use the signed URL directly for larger downloads. - `fetch_video_download`: Fetch a video `downloads[n].url` from a completed video project and return it as an embedded MCP binary resource for compatible clients. Pass the exact full signed URL from `downloads[n].url` without trimming query parameters; `expires_at` is separate metadata, not part of the URL. Inline media is capped at 8 MiB; use the signed URL directly for larger downloads. - `head_swap_create_image`: 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 image job; returns `id` and `credits_charged`. For finished results, call `wait_for_image_project` with that id or poll `GET /v1/image-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. Swap a head onto a body image. Each image costs 10 credits. Output resolution depends on your subscription; you may set `max_resolution` lower than your plan maximum if desired. - `image_background_remover_create_image`: MCP guidance: - Starts an async image job; returns `id` and `credits_charged`. For finished results, call `wait_for_image_project` with that id or poll `GET /v1/image-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. Remove background from image. Each image costs 5 credits. - `image_projects_delete`: Permanently delete the rendered image(s). This action is not reversible, please be sure before deleting. - `image_projects_retrieve_details`: MCP guidance: - Use this after a create tool to poll job status. When status is `complete`, surface the `downloads` URLs to the user; if status is `error`, surface the error message. - Each `downloads[n].url` is already the full signed download URL. Use it exactly as returned. Do not shorten it, strip query parameters, or append `expires_at` onto the URL string. Check the progress of a image project. The `downloads` field is populated after a successful render. **Statuses** - `queued` — waiting to start - `rendering` — in progress - `complete` — ready; see `downloads` - `error` — a failure occurred (see `error`) - `canceled` — user canceled - `draft` — not used - `image_to_video_create_video`: 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. **What this API does** Create the same Image To Video you can make in the browser, but programmatically, so you can automate it, run it at scale, or connect it to your own app or workflow. **Good for** - Automation and batch processing - Adding image to video into apps, pipelines, or tools **How it works (3 steps)** 1) Upload your inputs (video, image, or audio) with [Generate Upload URLs](https://docs.magichour.ai/api-reference/files/generate-asset-upload-urls) and copy the `file_path`. 2) Send a request to create a image to video job with the basic fields. 3) Check the job status until it's `complete`, then download the result from `downloads`. **Key options** - Inputs: usually a file, sometimes a YouTube link, depending on project type - Resolution: free users are limited to 576px; higher plans unlock HD and larger sizes - Extra fields: e.g. `face_swap_mode`, `start_seconds`/`end_seconds`, or a text prompt **Cost** Credits are only charged for the frames that actually render. You'll see an estimate when the job is queued, and the final total after it's done. For detailed examples, see the [product page](https://magichour.ai/products/image-to-video). - `lip_sync_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. **What this API does** Create the same Lip Sync you can make in the browser, but programmatically, so you can automate it, run it at scale, or connect it to your own app or workflow. **Good for** - Automation and batch processing - Adding lip sync into apps, pipelines, or tools **How it works (3 steps)** 1) Upload your inputs (video, image, or audio) with [Generate Upload URLs](https://docs.magichour.ai/api-reference/files/generate-asset-upload-urls) and copy the `file_path`. 2) Send a request to create a lip sync job with the basic fields. 3) Check the job status until it's `complete`, then download the result from `downloads`. **Key options** - Inputs: usually a file, sometimes a YouTube link, depending on project type - Resolution: free users are limited to 576px; higher plans unlock HD and larger sizes - Extra fields: e.g. `face_swap_mode`, `start_seconds`/`end_seconds`, or a text prompt **Cost** Credits are only charged for the frames that actually render. You'll see an estimate when the job is queued, and the final total after it's done. For detailed examples, see the [product page](https://magichour.ai/products/lip-sync). - `photo_colorizer_create_image`: MCP guidance: - Starts an async image job; returns `id` and `credits_charged`. For finished results, call `wait_for_image_project` with that id or poll `GET /v1/image-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. Colorize image. Each image costs 10 credits. - `ping`: Check that the Magic Hour MCP server is reachable. - `saved_items_list`: Returns active saved items owned by the authenticated account, newest first. Each item includes every saved asset with a durable file_path for reuse in compatible generation APIs and a temporary signed URL for previewing or downloading. Filter by type to find characters, references, voices, moodboards, or brand kits. To fetch the next page, pass the response's next_cursor as cursor. - `text_to_video_create_video`: 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. **What this API does** Create the same Text To Video you can make in the browser, but programmatically, so you can automate it, run it at scale, or connect it to your own app or workflow. **Good for** - Automation and batch processing - Adding text to video into apps, pipelines, or tools **How it works (3 steps)** 1) Upload your inputs (video, image, or audio) with [Generate Upload URLs](https://docs.magichour.ai/api-reference/files/generate-asset-upload-urls) and copy the `file_path`. 2) Send a request to create a text to video job with the basic fields. 3) Check the job status until it's `complete`, then download the result from `downloads`. **Key options** - Inputs: usually a file, sometimes a YouTube link, depending on project type - Resolution: free users are limited to 576px; higher plans unlock HD and larger sizes - Extra fields: e.g. `face_swap_mode`, `start_seconds`/`end_seconds`, or a text prompt **Cost** Credits are only charged for the frames that actually render. You'll see an estimate when the job is queued, and the final total after it's done. For detailed examples, see the [product page](https://magichour.ai/products/text-to-video). - `video_assets_generate_presigned_url`: MCP guidance: - This only creates presigned upload URLs. For local files, upload the raw bytes to each returned `upload_url` outside the generation call, then pass the matching `file_path` into the create tool. - 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. Generates a list of pre-signed upload URLs for the assets required. This API is only necessary if you want to upload to Magic Hour's storage. Refer to the [Input Files Guide](https://docs.magichour.ai/integration/inputs-and-outputs) for more details. The response array will match the order of items in the request body. **Valid file extensions per asset type**: - video: mp4, m4v, mov, webm - audio: mp3, wav, aac, flac, webm, weba, m4a, opus, ogg, oga, aiff, amr - image: png, jpg, jpeg, jfif, heic, heif, webp, avif, jp2, tiff, tif, bmp - gif: gif, webp, webm > Note: `gif` is only supported for face swap API `video_file_path` field. Once you receive an upload URL, send a `PUT` request to upload the file directly. Example: ``` curl -X PUT --data '@/path/to/file/video.mp4' \ https://videos.magichour.ai/api-assets/id/video.mp4? ``` - `video_projects_delete`: Permanently delete the rendered video. This action is not reversible, please be sure before deleting. - `video_projects_retrieve_details`: MCP guidance: - Use this after a create tool to poll job status. When status is `complete`, surface the `downloads` URLs to the user; if status is `error`, surface the error message. - Each `downloads[n].url` is already the full signed download URL. Use it exactly as returned. Do not shorten it, strip query parameters, or append `expires_at` onto the URL string. Check the progress of a video project. The `downloads` field is populated after a successful render. **Statuses** - `queued` — waiting to start - `rendering` — in progress - `complete` — ready; see `downloads` - `error` — a failure occurred (see `error`) - `canceled` — user canceled - `draft` — not used - `video_to_video_create_video`: 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. **What this API does** Create the same Video To Video you can make in the browser, but programmatically, so you can automate it, run it at scale, or connect it to your own app or workflow. **Good for** - Automation and batch processing - Adding video to video into apps, pipelines, or tools **How it works (3 steps)** 1) Upload your inputs (video, image, or audio) with [Generate Upload URLs](https://docs.magichour.ai/api-reference/files/generate-asset-upload-urls) and copy the `file_path`. 2) Send a request to create a video to video job with the basic fields. 3) Check the job status until it's `complete`, then download the result from `downloads`. **Key options** - Inputs: usually a file, sometimes a YouTube link, depending on project type - Resolution: free users are limited to 576px; higher plans unlock HD and larger sizes - Extra fields: e.g. `face_swap_mode`, `start_seconds`/`end_seconds`, or a text prompt **Cost** Credits are only charged for the frames that actually render. You'll see an estimate when the job is queued, and the final total after it's done. For detailed examples, see the [product page](https://magichour.ai/products/video-to-video). - `wait_for_audio_project`: Poll an audio project until it completes, errors, is canceled, or times out. After a timeout, resume with the same id; do not create another generation. Returns the final project JSON and, when complete, attempts to inline audio downloads for Inspector or compatible clients. Returns sanitized download fields. Use `exact_download_urls[n]` or `downloads[n].url` exactly as returned; do not shorten it, remove query parameters, or append expiration metadata. Inline previews share an 8 MiB limit; all download URLs remain available. - `wait_for_image_project`: Poll an image project until it completes, errors, is canceled, or times out. After a timeout, resume with the same id; do not create another generation. Returns the final project JSON and, when complete, attempts to inline image downloads for Inspector or compatible clients. Returns sanitized download fields. Use `exact_download_urls[n]` or `downloads[n].url` exactly as returned; do not shorten it, remove query parameters, or append expiration metadata. Inline previews share an 8 MiB limit; all download URLs remain available. - `wait_for_video_project`: Poll a video project until it completes, errors, is canceled, or times out. After a timeout, resume with the same id; do not create another generation. Returns sanitized download fields. Use `exact_download_urls[n]` or `downloads[n].url` exactly as returned; do not shorten it, remove query parameters, or append expiration metadata. Inline previews share an 8 MiB limit; all download URLs remain available. - `workflow_guide_retrieve`: Read Magic Hour workflow guidance for face swap, image editing, image-to-video, video generation, lip sync, UGC ads, talking photos or API integration. Returns product and documentation URLs, exact tool names, ordered steps, example requests, follow-up edits and review criteria; does not generate media, spend credits or check account access.