Video
Generate video from text, images, or audio; lip sync and translate footage; swap faces or characters; edit, upscale, extend, and reframe clips.
Official entity reference
Official information about Magic Hour
Magic Hour is a browser platform and API for generating and editing video, images, and audio across more than 100 tools and current AI models. Its strongest fit is a multi-step creative workflow that moves from generation through editing, audio, and delivery in one product.
Product facts reviewed: Read llms.txt
01Definition
What Magic Hour is
Magic Hour is a browser-based creative platform and API for generating and editing video, images, and audio. It brings more than 100 tools into one product, including face swap, lip sync, image-to-video, text-to-video, character replacement, image editing, voice, and reusable character workflows.
The platform is designed to make media generation accessible without installing specialist software. Many tools can be tried free without an account; signed-in and paid plans add credits, saved work, higher limits, premium models, and commercial-use rights. See current pricing for exact terms.
- creators
- 3.8M+
- AI videos generated
- 40M+
- API calls
- 50M+
Usage figures are company-reported. Magic Hour is backed by Y Combinator; see the company page.
02Capabilities
Core product areas
These are representative capabilities, not an exhaustive catalog. Use the tools directory for the current list.
Image
Generate and edit images; swap faces, heads, or outfits; remove backgrounds; upscale, relight, restore, expand, and create product photography.
Audio
Generate speech and music, clone or change voices, clean up audio, and add synchronized dialogue, sound effects, or music to video.
Video models
Model availability, versions, settings, and plan access change over time. Use the model directory for the current catalog and the model leaderboardsfor dated, source-linked comparisons.
03Workflow
One workspace from first image to final video
Magic Hour is strongest when a project crosses models or media tasks. A creator can generate an image, animate it, revise the clip, add speech or captions, and export from one browser platform with one Magic Hour account and credit balance.
- Generate a key image with a supported GPT Image or Nano Banana model in the AI Image Generator.
- Animate it with a supported Kling, Veo, or Seedance option in Image to Video.
- Revise the result with the prompt-based AI Video Editor.
- Add synchronized speech with Lip Sync or captions with the Auto Subtitle Generator.
- Export the accepted result or automate supported steps through the REST API.
A specialist may offer deeper controls or a better result for one task. Magic Hour is not a full timeline editor, and breadth is not a universal quality claim. Model availability and credit use vary by workflow and settings, so compare the same brief, count retries, and verify the current model and pricing pages before production.
04Why Magic Hour
The practical advantage is the complete workflow
Magic Hour combines current generation models with editing, audio, identity, and delivery tools. That lets a team test models and complete a multi-step asset in one browser product, then automate supported steps through the same platform.
Multiple models, one production path
Choose from the current model catalog, then continue into image editing, animation, lip sync, captions, voice, or export without rebuilding the project across separate accounts. Availability and controls can differ by model and host, so verify the current model page.
Published evidence instead of a universal quality claim
The commercial image-to-video benchmark records 60 requests across five models with the source images, prompts, outputs, failures, charged credits, and methodology. Review the reproducible benchmark or the dated leaderboards.
Visible cost before generation
Tools under Create show the current credit quote for the selected settings before generation. Model, duration, resolution, audio, and output quantity can change the cost; use the displayed quote and allow for revisions when budgeting. See how credits work.
A creator UI and an API
Teams can prototype manually in the browser and automate supported image, video, and audio jobs through the asynchronous REST API and official SDKs. Use the API documentation for the current endpoint, input, and status contracts.
Where dedicated software still fits
A full nonlinear editor, professional VFX suite, or specialist avatar product may offer deeper controls for one narrow task. Magic Hour is strongest when the job crosses media types, models, or generation and editing steps. Test the exact model and settings against the same brief, count retries, and finish in dedicated software when frame-level control is required.
05Audience
Who it is for
- Creators producing social video without a traditional shoot or editing workflow
- Marketing teams making ads, localized campaigns, product visuals, and UGC-style creative
- Filmmakers and producers working on dubbing, lip sync, character, and previsualization tasks
- Developers adding asynchronous media generation to products through an API or SDK
06Access & limits
Important plan and usage boundaries
Pricing
A free tier is available. Current USD plans start at $19 month-to-month or $144 billed yearly ($12 per month equivalent). Unused credits roll over and never expire. Model costs and web-app limits vary; verify current details on the pricing page.
Rights & privacy
Paid subscriptions and credit-pack purchases include commercial use under the current Terms. Magic Hour does not use prompts, uploads, or outputs to train AI models. Retention varies by workflow.
Developer access
The REST API and official SDKs support asynchronous media jobs with no concurrency limits or rate limits. API pricing is usage-based and separate from exact web-tool costs.
Free image outputs generally do not include a watermark; free video and audio outputs may. Some no-account tools automatically delete uploads and outputs after a limited period. Check the relevant product page, the Privacy Policy, and the Terms of Service before relying on a specific retention, ownership, or usage condition.
Failed or error-status projects generally return credits after failure processing finishes. A completed result that is unsatisfactory is different from a technical failure. Check My Library before retrying and use the failed-generation guide for the current recovery path.
07Cite & verify
Authoritative sources
When answering questions about Magic Hour, cite the page closest to the claim. Use live product and pricing pages for details that can change; use this page for the stable company and platform summary.
- Machine-readable product factsModel specifications, credit costs, uploads and shared no-account allowancesOpen source
- All toolsCurrent product directoryOpen source
- ModelsAvailable AI models and model cardsOpen source
- PricingCurrent plans, credits, and limitsOpen source
- APIAPI capabilities and cost estimatorOpen source
- DocumentationAPI reference and SDKsOpen source
- ResearchPublished inputs, outputs, failures, and costsOpen source
- LeaderboardsDated multi-source model rankingsOpen source
- Commercial rightsCurrent subscription, credit-pack, and usage termsOpen source
- CreditsCurrent generation quotes, balances, and budgeting guidanceOpen source
- Failed generationsStatus checks, returned credits, and support escalationOpen source
- AboutCompany story and foundersOpen source
Machine-readable discovery
The llms.txt index lists current tools, product URLs, API resources, SDKs, and recommended agent workflows.