Magic Hour Changelog
November 2026
Use Your Original Audio with Wan 3.0
Wan 3.0 is now available in Talking Photo and Audio-to-Video. Add a photo and a recording, then choose a 6- to 10-second section to animate at 720p. The export keeps that section of your original recording as its soundtrack.
Use Talking Photo to animate a portrait with your own voice, or open Audio-to-Video to guide the motion with a photo and a clip of your recording.
Try Wan 3.0 in Audio-to-Video
Choose Wan 3.0 and add a photo and audio.
Choose Which Part of a Song to Preview
The free Music Video Generator now lets you choose any five seconds of your song for a preview. It used to start at the beginning of the track.
Upload a song and drag the highlighted range to the part you want, such as the chorus. The free preview is 480p.

Try the Free Music Video Preview
Upload your song and choose the section to preview.
Write Longer Prompts for LTX 2.5 and Seedance
In the dashboard, Text-to-Video and Image-to-Video now accept up to 4,000 characters for LTX 2.5, Seedance 2.0 and Seedance 2.5, up from 2,000. LTX 2.5 also supports the higher limit in both API endpoints.
Use the extra space to describe the order of actions, camera moves, and details you want to keep consistent across a clip.
Try a Longer Video Prompt
Open Text-to-Video with LTX 2.5 selected.
Nano Banana 2 Is Now Nano Banana 2.1 in Magic Hour
Nano Banana 2 has been retired as a separate option in Magic Hour. New image generations and edits that previously used Nano Banana 2 now run on Nano Banana 2.1.
Existing Image Generator and Image Editor API integrations can keep sending the old model value. If you are updating your integration, use nano-banana-2.1 to make the model explicit.
View Nano Banana 2.1
See the model for image generation and editing.
Generate and Edit Images with Seedream 5 Flash
Seedream 5 Flash is now available on paid plans in AI Image Generator and AI Image Editor. Generate images from a prompt or edit an image using additional images as references. Outputs go up to 2K.
Seedream 5 Flash is also available through the Image Generator and Image Editor APIs. Set model to seedream-v5-flash.
Use it when you want to try several concepts before refining one. You can generate up to four images at a time in AI Image Generator, then open your favorite in AI Image Editor and use another image to guide the edit.
Prompt used for the example below: “Triptych of a ceramic coffee mug: a close-up of the handle, the mug on a breakfast table, and the mug held in two hands, consistent warm style”.

Try Seedream 5 Flash
Choose Seedream 5 Flash and describe the image you want.
Edit Videos with Kling 4.0 in the API
Kling 4.0 is now available to paid users of the AI Video Editor API. Set the model to kling-4.0 to edit 3- to 30-second clips at 720p or 1080p.
Send the source clip and describe what you want to change. You can also attach named image references and mention them with @name in your prompt.
View the Video Editor API Docs
See the request fields for Kling 4.0.
Use Google Omni 1.2 in the Video API
Google Omni 1.2 is available in the Image-to-Video and AI Video Editor API endpoints. Use model gemini-omni-1.2 for clips from 3 to 15 seconds.
In Image-to-Video, set audio to false when you want a silent clip for a separate soundtrack. For example, animate a product photo and add the narration later.
View the Image-to-Video API Docs
See the model and audio request fields.
Create Longer Videos with Google Omni 1.2
Google Omni 1.2 replaces 1.1 in the web app. Use it to create videos from text or images, or edit existing clips. Those modes support clips from 3 to 15 seconds.
Check the credit estimate before generating. The cost depends on resolution, duration, and whether you use a video reference.
Google’s release demo shows a beehive explainer with labels that follow the action.
Watch Google’s Omni 1.2 beehive demo
Try Google Omni 1.2
Choose a duration and describe your clip.
Write Longer Prompts for Kling 4.0
Kling 4.0 now accepts prompts up to 10,000 characters in Text-to-Video and Image-to-Video, up from 2,500.
That gives you room to describe a sequence in detail without trimming the instructions that matter. For example, spell out the camera movement, what the subject does, and how the lighting changes from the opening shot to the end.
Try Kling 4.0
Open Kling 4.0 in Text-to-Video.
Use More Seedance 2.5 References in the Text-to-Video API
Seedance 2.5 now accepts up to 50 named references per Text-to-Video API request, up from 10. You can include up to 30 images, 10 videos, and 10 audio clips within that total.
Mention each reference by name with @ in your prompt. Audio references need an image or video reference. The worker trims video and audio references to 30 seconds combined for each media type.
For example, use 12 product images and a short motion reference to guide one clip.
View Text-to-Video API Docs
See the references field and example request.
Export a Video Frame as a JPEG
You can now save a still from a completed video directly in My Library. Open the video, choose Export frame (JPEG) from the More menu, and scrub to the moment you want.
Select Download JPEG to save the frame without using generation credits. For example, pull a still for a thumbnail or use it as a starting image for your next clip.
Open My Library
Choose a finished video and pick a frame.
Save and Reuse a Generation Setup
Save a completed generation as a workflow to keep its prompt, model, settings, and inputs together in My Library.
For example, if you want to make another version of an image edit, choose Save workflow on the finished result. Then go to My Library, open Saved Workflows, and choose Use workflow. Your setup opens in the creator so you can make changes before generating again.

Open Saved Workflows
Find and reopen your saved setups.
Generate Images from Text with Seedream 4.5
Seedream 4.5 can now create images from a text prompt in AI Image Generator. Creator, Pro, and Business users can generate up to four images per request at resolutions from 640px to 4K.
For example, start with no reference image and enter “A red ceramic teapot on a walnut table beside a sunlit window, editorial product photograph, natural shadows, no text.” The image below was generated with that prompt at 640px in a 9:16 aspect ratio.

Try Seedream 4.5
Describe the image you want to create.
A 256,000-Credit Pack for Larger Projects
You can now buy 256,000 credits in one purchase. The new pack costs $480 USD before applicable taxes.
If you are budgeting for a larger batch of videos or images, choose the 256,000-credit option under Credit Packs and review the total in your currency.
View Credit Packs
Compare one-time credit packs.
Create Nano Banana 2.1 Storyboards in 4, 9, or 16 Panels
Nano Banana 2.1 is now available in AI Storyboard Generator on all plans. Choose 4, 9, or 16 panels, use a 16:9 or square layout, and add up to 10 reference images. Each storyboard costs 150 credits.
Plan a sequence of shots from one scene description. The model uses its 4K setting for the full storyboard grid; individual panel resolution varies with the layout.
Example: four shots of a red-haired girl in a yellow raincoat and her corgi exploring a lighthouse, climbing the stairs, discovering a map, and watching a storm. These are four panels from one Nano Banana 2.1 storyboard.

Build a Storyboard
Choose Nano Banana 2.1 in the model menu.
Use Characters, Moodboards, and Brand Kits on Free Accounts
Signed-in free accounts can now use saved Characters, Moodboards, and Brand Kits with Qwen Image 2.1 to create and edit images. You can also use compatible Qwen-generated start frames in LTX video workflows. Normal generation credits still apply.
Reuse your saved visual references without a paid subscription. Qwen accepts up to three input images in total, including uploads and every image in a Brand Kit. Exact text and logos can vary.
For example, use a candle Brand Kit to create a product photograph featuring its amber jar and botanical label, as shown below.

Create with Saved References
Choose Qwen Image 2.1 and add your saved references.
Higher Credit Rates for H3 Character Replace and Animate
MiniMax H3 in Character Replace and Animate now costs 48 credits per second at 480p, 96 at 720p, and 144 at 1080p. These rates are double the previous rates.
Review the generation quote when planning your clip. 480p is available across plans; 720p and 1080p require Creator, Pro, or Business in the dashboard.
For example, a five-second H3 character replacement at 480p now costs 240 credits.
Review H3 Character Pricing
Choose MiniMax H3 and review the quote.
Edit Existing Videos with Kling 4.0
Kling 4.0 is now available in AI Video Editor on Creator, Pro, and Business. Upload a video, describe the change, and select a clip from 3 to 30 whole seconds. The edit keeps the source video’s aspect ratio.
Choose 720p at 96 credits per second or 1080p at 144 credits per second. Use text instructions to change an object, color, or scene without rebuilding the clip from scratch.
For example, upload a street scene and ask: “Make all the yellow taxis bright green.” A five-second edit at 720p costs 480 credits.
Edit a Video with Kling 4.0
Upload a clip and describe your edit.
Longer Voiceovers and More Control through the API
AI Voice Generator and Voice Cloner API requests now accept scripts up to 5,000 characters. Both support language selection and speech speed from 0.75× to 1.30×.
Build longer narrations and set the delivery pace directly in your request. Use GET /v1/voices to find curated voice IDs and their preview URLs. API key and credit requirements apply.
For example, choose a voice ID from the catalog and request a Russian voiceover with style.language set to russian and style.speed set to 0.9. Keep the returned project ID to check its progress.
View Voice API Docs
See request fields and examples.
MiniMax H3 for Talking Photos
You can now choose MiniMax H3 in the Talking Photo dashboard. Add a photo and speech, with an optional prompt to guide the animation.
Use a short script to turn a portrait into a talking video. H3 supports clips up to 15 seconds: 480p is available across plans, and 720p requires Creator, Pro, or Business. This mode does not support 1080p.
For example, upload a portrait and a five-second voiceover, choose MiniMax H3 under Generation Mode, and add a brief motion prompt.
Try H3 Talking Photo
Choose MiniMax H3 under Generation Mode.
New Credit Rates for Kling 3.0
Kling 3.0 generation now uses updated credit rates in Text-to-Video, Image-to-Video, and their API workflows.
For 720p clips without a video reference, silent video costs 60 credits per second, up from 48. Video with audio costs 84 credits per second, up from 72. Other resolutions and reference modes have different rates.
For example, a five-second 720p clip costs 300 credits without audio or 420 credits with audio. Review the displayed quote before starting a generation.
Check Kling 3.0 Quote
Choose your duration, resolution, and audio setting.
Choose Exact Clip Durations with Wan 3.0
Wan 3.0 now supports every whole-second duration from 2 to 30 seconds in Text-to-Video and Image-to-Video, including lengths such as 12 or 18 seconds.
Set a clip length that fits your edit instead of choosing a nearby preset. Wan 3.0 requires Creator, Pro, or Business in the dashboard; credit costs still depend on resolution and duration.
For example, choose Wan 3.0 and set 12 seconds for a short scene, or use 18 seconds when animating a starting image.
Create with Wan 3.0
Choose Wan 3.0 and set your clip length.
October 2026
Eight New Spanish and Portuguese Voice Presets
AI Voice Generator now offers eight additional voice presets: Fernanda, Gabriela, Regina, and Valeria for Mexican Spanish; Florencia and Lautaro for Argentine Spanish; Adriana and Davi for Brazilian Portuguese.
Audition their samples to choose a voice suited to your audience before generating speech. Search by name in the voice picker to find each preset.
For example, preview Florencia for an Argentine Spanish script or Davi for Brazilian Portuguese, then select the voice for your script.
Explore the New Voices
Preview the new voices and choose one for your script.
MiniMax H3 for Video Extension and Sound
MiniMax H3 is now a model option in the dashboard’s AI Video Extender and Video to Audio tools. Use it to continue an existing shot or generate a separate audio track for a video.
In Video to Audio, H3 supports selected clips up to 15 seconds. 480p is available across plans; 720p and 1080p require a Creator, Pro, or Business subscription.
For example, choose MiniMax H3 to extend a shot by five seconds, or describe the sound for a three-second scene in Video to Audio.
Try H3 Video Extender
Choose MiniMax H3 in the model menu.
Create Looping GIFs with LTX 2.5
AI GIF Generator now uses LTX 2.5 to turn a text prompt into a looping GIF. In the full creation tool, choose a duration from 2 to 10 seconds; generation costs 24 credits per second.
The free tool creates three-second square GIFs without a watermark, under the existing shared daily limit. GIFs export without sound at 12 frames per second, with the longer side capped at 448 pixels.
For example, describe a waving mascot or a small looping animation for a chat. In the full tool, choose four seconds for a short reaction GIF, then download the result to share.
Create a Looping GIF
Try the free three-second GIF tool.
Create Videos with Kling 4.0
Kling 4.0 is now available in Text-to-Video and Image-to-Video on Creator, Pro, and Business subscriptions. Create clips from 3 to 30 seconds at 720p or 1080p, with optional audio.
Guide the scene with up to 10 image references and one video reference, or use start/end frames. Reference inputs and start/end frames are separate workflows and cannot be combined in one request.
For example, add two character images as references and describe a scene in which they interact. Select Kling 4.0 in the model menu before generating.
Watch the official Magic Hour review below for Kling 4.0 video examples.
Try Kling 4.0
Choose Kling 4.0 in the model menu.
Cloudflare Workflows and Inngest Examples
New first-party examples let you run Magic Hour image generation and image-to-video inside Cloudflare Workflows or Inngest. They include accepted-project checkpoints, status checks, waiting steps, and recovery using a saved project ID.
Start with the Cloudflare Workflows example or the Inngest example. Run the code in your own Node.js development environment and configure deployment for your application.
For example, generate a product image, then use its completed output as the input to an image-to-video workflow. Keep the accepted project ID when resuming so you can check that job instead of submitting it again.
You need a Magic Hour API key and API credits. Download files you want to retain before their temporary result URLs expire.
Get a Magic Hour API Key
Connect the workflow examples to your Magic Hour account.
New Integrations for Apps and Automation
Add Magic Hour generation to your apps and automation workflows with new first-party adapters and ready-to-use flows.
- AI SDK, October 23: our GitHub-distributed AI SDK provider supports image generation and experimental image-to-video with AI SDK 7 on Node.js 22+.
- Node-RED, October 23: import the image workflow to create an image job, then manually run Recover to check its project ID and save the completed PNG to your configured folder.
- Activepieces, October 24: install our custom piece version 0.1.1 in self-hosted Community Edition to add Magic Hour generation and project-status actions. This is not a built-in Cloud catalogue integration.
- LiteLLM, October 25: use the first-party Python adapter from its GitHub release wheel to generate one square Z-Image Turbo image per request, returned as a download URL.
These integrations require a Magic Hour API key and API credits. Download links are temporary, so save files you want to keep.
For example, use Create in the Node-RED flow to submit an image job, keep its project ID, then use Recover after completion to save the PNG without submitting another generation.
Get a Magic Hour API Key
Connect your chosen integration to your Magic Hour account.
Create Images with Strands Agents
You can now create and retrieve Magic Hour images from a Python application using the Strands Agents MCP client.
The integration guide covers submitting a prompt, saving the project ID, waiting for completion, and downloading the result. Reuse that ID to retrieve an accepted job instead of starting another generation. A Magic Hour API key and sufficient credits are required.
This product image was created through the Strands client using this prompt: “Studio product photograph of an original unbranded cobalt-blue reusable water bottle on a pale stone pedestal, sunrise desert backdrop, no people, text, or logos.”

Use Magic Hour with Strands
Follow the Python integration guide.
Create Longer Music Videos and Scenes
The full creation tools now accept longer clips and timelines, so you can work with longer songs, performances, and multi-scene sequences.
- AI Music Video Generator: clips up to 6 minutes, up from 5 minutes.
- Image-to-Video Keyframes: target timelines up to 3 minutes with LTX 2.5 or MiniMax H3. The 60-second per-segment limit is unchanged.
- AI Video Editor with LTX 2.5: clips up to 60 seconds at 480p or 720p, and 30 seconds at 1080p, in the web tool and API.
- Expressive Talking Photo: clips up to 60 seconds. Realistic mode's existing limit is unchanged.
- Audio-to-Video: uploaded audio clips up to 60 seconds.
- Video Extender with LTX 2.5: new 15-second and 20-second extension choices.
Your account's duration and resolution limits still apply. Free-tool allowances are unchanged.
For example, choose a music-video trim of up to six minutes, or lay out a three-minute keyframe sequence using shorter transitions between images.
Open Music Video Generator
Add your song and choose the part to turn into a video.
Use MiniMax H3 for Character Replace and Animate
MiniMax H3 is now an option in the full Character Replace tool and the Character Replace API. Wan Animate remains the default.
Choose Replace to swap a person in a driving video, or Animate to move your character image using the video's motion. Both modes accept an optional prompt.
H3 supports clips up to 60 seconds, subject to your account's duration limits. 720p and 1080p require paid-plan access.
For example, select MiniMax H3, choose Animate image, and add a character image plus a driving video. Use the optional prompt to give the model additional guidance.
Open Character Replace
Select MiniMax H3 in the model menu.
Voice Preset Availability Update
Dove Cameron and Davido have been removed from the Voice Generator preset library. Choose another available voice before running a saved workflow that uses either preset.
API integrations must also replace these values in style.voice_name on POST /v1/ai-voice-generator. Requests using a removed value no longer pass validation.
For example, update a saved request that used Davido by replacing its style.voice_name with an available preset; your script can stay unchanged.
Browse Available Voices
Choose a replacement from the current Voice Generator library.
Create Magic Hour Videos inside Framer
Magic Hour Video is now available in the Framer Marketplace. Select an image in your design, describe its motion, preview the generated clip, and add it as a native Framer Video component without leaving the editor.
Use existing artwork to create motion for your site while keeping the original image in place. The plugin is free; generation requires a Magic Hour account, API key, and sufficient API credits.
For example, select a product image, generate a short motion clip, then choose Add video to design. The inserted video uses muted looping playback; a seamless visual loop is not guaranteed.
Open Magic Hour Video in Framer
Install the plugin from its Framer Marketplace listing.
Browse Meme Formats and Start with a Template
Browse 100 meme formats in the new Meme Templates directory. Each format has its own page with an explanation, an example idea, captions, and the matching generator already selected.
Find a format that fits your joke before writing the idea, without searching through the generator's full template list.
For example, open Drake Hotline Bling to see how its reject-and-approve layout works, then enter your own idea in the selected template's generator.
Browse Meme Templates
Find a format for your next meme.
More Voice Presets for Audio Creation
The dashboard's AI Voice Generator, Voice Changer, and Voice Cloner now include an expanded catalog of curated voice presets. New Voice Generator and Voice Changer sessions start with Ashley selected.
Explore more voices directly in the tools when choosing an identity for narration or a voice workflow.
For example, open the voice picker and compare Ashley, Wade, Lauren, or Nicholas before making your selection. The expanded dashboard catalog is separate from the public Text to Speech page's voice selection.
Explore Voice Presets
Find your next narration voice in the dashboard.
Draft and Revise UGC Scripts in the Editor
Generate Script is now available in the signed-in UGC Ad Generator. Choose an actor or opening image, then generate a script or revise an existing draft with your selected duration in mind.
Develop the words alongside the actor and product instead of moving between writing and video tools. You can regenerate the script or undo the change before creating a video.
For example, select your actor, add a short product idea or draft, and choose Generate Script. Review the returned text before generating. This works in text-script mode, not uploaded-audio mode.
Draft a UGC Script
Choose an actor and review your script before generating.
Upscale Videos through the API
Video Upscaler is now available through the API for paid plans. Choose a target resolution of 720p, 1080p, 2K, or 4K while preserving the source video's aspect ratio.
Upscale clips inside your own application or video-processing workflow without manually moving between tools. Processing uses Magic Hour credits and supports up to 4,000 frames at a maximum of 30 fps.
For example, upload a 720p clip and request a five-second segment at 1080p with POST /v1/ai-video-upscaler. Use the returned project ID to check completion and retrieve the download. Choose a target above your source resolution: a clip already at the target resolution may be returned unchanged and still use credits.
View Video Upscaler API docs
Read the endpoint, request parameters, and examples.
Generate Music through the API
The Music Generator API is now available. Describe a track, choose a duration from 15 to 300 seconds, and request instrumental music or provide lyrics.
Add music generation to your own app or automated video workflow using your Magic Hour API key and credits.
For example, request a 30-second instrumental track with the prompt "Warm acoustic guitar for a product walkthrough." Submit it to POST /v1/ai-music-generator, then check GET /v1/audio-projects/{id} for completion and download links.
View Music Generator API docs
Read the endpoint, request parameters, and examples.
Video Expander Pricing Now Follows Output Resolution
Video Expander now prices new generations at 24 credits per second for 480p, 48 for 720p, and 72 for 1080p. Previously, all three resolutions used the 480p rate.
Higher-resolution expansions now cost more credits. Auto uses the resolved output resolution. Existing jobs are not repriced, and resolution access by plan is unchanged.
For a five-second clip, that means 120 credits at 480p, 240 at 720p, or 360 at 1080p. Review the displayed credit cost before generating.
Open Video Expander
Choose an output resolution and review its credit cost.
Colorize Videos up to Five Minutes
Video Colorizer now supports clips up to five minutes in the full creation flow, up from 45 seconds. The free-tool limit remains 10 seconds.
Colorize longer scenes without first splitting them into short clips.
For example, submit a one-minute black-and-white clip as a single Colorizer job, then download the completed video.
Open Video Colorizer
Add your clip and review the duration and credit cost.
Guide Google Omni Text-to-Video with Image References
Google Omni in Text-to-Video now accepts an image reference directly. Use a photo as visual guidance while describing a new scene.
This workflow requires a Creator, Pro, or Business subscription in the dashboard. Image-reference generation is separate from Start & End Frame mode and does not support audio references.
For example, add a product photo and describe a new setting and camera movement. Select Google Omni, stay in Prompt mode, and add the photo under Add References.
Try Google Omni References
Add a photo under Add References in Prompt mode.
Choose FLUX.2 Klein for Image Body Swaps
Image Body Swap now lets you choose between Nano Banana and FLUX.2 Klein. Nano Banana remains the default.
Choose a model and supported output resolution to fit your body-swap workflow. FLUX.2 Klein offers 480p, 720p, 1080p, and 1440p settings; credit cost depends on the resolution.
For example, choose FLUX.2 Klein with 720p for an image body swap, or select 1080p when you need a larger output. Review the displayed credit cost before generating.
Choose a Body Swap Model
Compare model options and review the credit cost.
Build MiniMax H3 Videos with Multiple Keyframes
MiniMax H3 is now available in Keyframes mode. Add multiple images to guide a video through an ordered sequence, rather than providing only its first and last frames.
Set the sequence and segment durations while H3 prepares the transition prompts. This workflow does not offer editable per-keyframe prompts.
Example: arrange three images as two transitions and generate the sequence.
Try H3 Keyframes
Open the tool to try the released workflow.
Reuse Characters in Storyboards, UGC Ads, and Photo Tools
Saved Characters are now available in Storyboard's Add Image flow and through explicit Character selection for UGC Actor, Clothes Changer's Person Image, and Head Swap's Head input.
Reuse a saved identity in these inputs without finding and uploading its image again. The selected Character fills the relevant input; you can review the setup before generating.
Example: choose a saved Character as your UGC Actor and review its image in the Actor input and Opening Shot.
Use a Character in UGC Ads
Open the tool to try the released workflow.
Compare Your Upscaled Video with the Original
Completed Video Upscaler results now support a before/after slider and a side-by-side comparison view.
Inspect the result against the original without opening separate players. Comparison is available when the selected trim begins at the start of the source video, keeping both sides synchronized.
Example: open a supported completed upscale and switch between the comparison slider and side-by-side view.
Open Video Upscaler
Open the tool to try the released workflow.
Give Your Characters a Saved or Recorded Voice
Character creation and editing now let you choose a saved voice, record a voice, or upload voice audio without a preview setting. The same controls are available when adding a Character from AI Avatar.
Keep a reusable visual identity and voice together instead of setting up the voice again for each Character. New celebrity-preset selection is removed; existing Characters keep their previously saved voice when you leave it unchanged.
Example: edit a Character, choose a voice from My Voices or record one, then save.
Add a Voice to Your Character
Open the tool to try the released workflow.
Choose Seedance 2.0 Fast for Video Creation
Seedance 2.0 Fast is now a distinct option in Text-to-Video, Image-to-Video, and the public API. It is available on paid tiers at 480p or 720p for clips from 4 to 15 seconds.
Choose Fast explicitly rather than treating it as an interchangeable fallback. Supported workflows include audio, end frames, and media references; review the displayed credit quote for your settings.
Example: choose Seedance 2.0 Fast in the model picker and set your clip length and resolution. Release checks verified its model selection and request routing. The first-party Seedance family showcase below is not a measured Fast speed or output-quality comparison.

Try Seedance 2.0 Fast
Open the tool to try the released workflow.
Extend or Expand a Completed Video without Re-uploading
Completed dashboard video outputs now offer Extend Video and Expand Video in their menus, including supported Library detail views and mobile result menus.
Open the next tool with that output already loaded instead of downloading and uploading it again. Choose your trim, aspect ratio, and other settings before clicking Generate; opening the shortcut does not start a generation.
Example: open a completed video's menu, choose Expand Video, and review the prefilled clip. Release checks verified this handoff on desktop and mobile. The first-party illustration below shows Video Expander, not the menu. Image and audio output menus are unchanged.

Open Your Video Library
Open the tool to try the released workflow.
Choose Talking Photo API Resolution Explicitly
The AI Talking Photo API now accepts an explicit resolution value of 480p or 720p. The older max_resolution field is deprecated but remains compatible.
Specify the output resolution directly in new integrations. When both fields are supplied, resolution takes precedence; supported model and plan limits still apply.
Example: set resolution to 720p in a new request. Existing integrations can keep using max_resolution while migrating.
View Talking Photo API Contract
Open the tool to try the released workflow.
Create UGC Ads Through the API
The public API now includes POST /v1/ai-ugc-ad. Supply a prepared image in assets.image_file_path and a script to create a UGC video using supported Kling 3.0, LTX 2.5, or Veo 3.1 options.
Integrate this prepared-image workflow without generating and polling a separate opening image. The endpoint returns a standard video project ID and uses the existing status and download endpoints; plan, duration, resolution, and credit limits still apply.
Example: send a prepared image and script to POST /v1/ai-ugc-ad, then use the returned project ID to check status and download the video.
View UGC API Contract
Open the tool to try the released workflow.
Set First and Last Frames with MiniMax H3
MiniMax H3 now accepts both a first and last image in Image-to-Video at 480p, 720p, and 1080p.
Choose where the clip begins and ends instead of guiding it with only a starting frame. Add a motion prompt to describe the transition between the two images.
Example: select MiniMax H3 in Start & End Frame mode and add both images. Release checks verified the uploads at 480p and completed worker output checks. The existing first-party H3 example below is not that test output.

Try H3 Start and End Frames
Open the tool to try the released workflow.
Video Upscaler v2 with a Free Short Preview
Video Upscaler v2 is now available in Magic Hour, with a short free preview on the public tool page and credit-based 2x upscaling available to Free-plan accounts.
Try the first three seconds before committing to a longer upscale. The free preview uses 2x scaling, is capped at 2K, and has a three-per-day allowance; longer generations use credits and applicable limits.
Example: upload a clip and review its short upscale preview before choosing a longer run. The existing first-party example below illustrates video upscaling, not a measured v2 quality or speed comparison.

Preview Video Upscaler
Open the tool to try the released workflow.
Record Audio for Talking Photo and Lip Sync
The public Talking Photo and Lip Sync audio pickers now include Record audio alongside uploads, text-to-speech, and sample audio.
Speak your own script without making a separate audio file first. Your selected photo or video stays in place, and cancelling the recorder keeps the existing input.
Example: open Add Audio, choose Record audio, allow microphone access, and finish your recording. Release checks verified that recorded WebM audio can produce a Talking Photo result. The first-party demo below illustrates Talking Photo, not that specific recording.

Record a Talking Photo Script
Open the tool to try the released workflow.
Guide Wan 3.0 with Native Media References
Wan 3.0 video creation now sends supported image, video, and audio references directly to the model instead of first synthesizing a hidden starting image.
Guide a new clip with the media you selected. Use native references as a separate workflow from first/last-frame generation; the two input modes cannot be combined.
Example: choose Wan 3.0, add reference media, and describe the clip you want. Release checks produced completed native-reference videos through both supported providers. The first-party video illustration below is not a Wan reference-quality benchmark.

Try Wan 3.0 References
Open the tool to try the released workflow.
Enhance Video Prompts with Your Starting Image
Signed-in users can now Enhance supported Text-to-Video prompts and LTX 2.5 Image-to-Video prompts without a preview setting. Image-to-Video uses the uploaded frame as context and can suggest motion from a blank prompt.
Review the suggested direction before generating, or choose Undo to restore your original wording. Existing model and sign-in restrictions still apply.
Example: upload a starting frame, choose Enhance, and review the suggested motion.
Enhance a Video Prompt
Open the tool to try the released workflow.
Generate Images with the Kestra Blueprint
An image-generation blueprint is available in Kestra's official blueprint repository. Use it to run Magic Hour image generation inside your Kestra workflows without a custom plugin.
Supply a prompt and aspect ratio. The workflow creates a job, checks that same project's status, and returns its download URL when complete.
For example, run the blueprint with a square product-image prompt, then download the finished image from the returned link for the next step in your pipeline.
You need a Magic Hour API key, API credits, and a running Kestra instance. The blueprint returns a temporary URL, not a permanently saved file. Add a download or storage step when you need to retain the image.
Open the Kestra Blueprint
Import the published image-generation workflow.
Nano Banana 2.1 for Image Generation and Editing
Nano Banana 2.1 is now available as a distinct model in AI Image Generator, AI Image Editor, and the public API using nano-banana-2.1.
Choose the new model for your image workflow while existing Nano Banana 2 API and template callers remain supported. Review the credit quote for your selected resolution and image count before generating.
For example, select Nano Banana 2.1 for a new product image or to edit an existing photo. The Magic Hour walkthrough below compares the two Nano Banana versions using image examples.
Try Nano Banana 2.1
Open the tool to try the released workflow.
Use Photo References with Qwen Image 2.1
Qwen Image 2.1 now accepts up to three primary reference photos in AI Image Generator and AI Image Editor. Free-plan users can use this supported reference workflow with credits.
Guide a creation with saved Characters or reference photos without a separate reference-feature subscription gate. Model access and credit costs still apply; Qwen Brand Kits are not supported.
Example: choose Qwen Image 2.1 and add reference photos to guide your prompt. The existing first-party Qwen showcase below illustrates image editing, not a three-reference quality benchmark.

Try Qwen Photo References
Open the tool to try the released workflow.
Enhance Image Prompts Before You Generate
Signed-in users can now use Enhance on supported AI Image Generator models and single-image AI Image Editor workflows without a preview setting.
Review a revised prompt before generating, and use Undo to return to your original wording. Guest, unsupported-model, multi-image, and video restrictions still apply.
Example: write a product-photo direction, choose Enhance, review the revised text, and Undo if you prefer your original.
Enhance an Image Prompt
Open the tool to try the released workflow.
Use a Video Reference with Kling 3.0
Kling 3.0 now supports one video reference, optionally combined with up to four image references, in its supported paid reference workflow.
Guide a new clip with an existing video and supporting images. Supported reference clips can retain or mute their original sound; this does not add newly generated audio.
Example: select Kling 3.0, add a video reference and an image, and describe your clip. Release evidence includes a completed three-second 720p reference generation with original sound. The existing first-party Kling showcase below is not that test output.

Try Kling Video References
Open the tool to try the released workflow.
Save Completed Image Outputs as Characters
Completed still-image results now offer Save as… in their menu. Name the image in the Save as Character dialog to add it to your Characters picker.
Reuse an image you already created as a Character without limiting this action to AI Influencer results. The action is not available for GIF, video, audio, processing, or failed outputs; each save creates a new Character.
Example: open a completed image's menu, choose Save as…, enter a name, and save. Release checks verified that the saved Character appears in the Image Generator picker on desktop and mobile. The first-party Characters showcase below illustrates reusable Characters.

Open Your Image Library
Open the tool to try the released workflow.
Restore More Aspect Ratios in Image Creation and Editing
AI Image Editor and AI Image Generator offer 4:3, 3:2, 4:5, and 2:3 alongside 16:9, 9:16, and 1:1. Auto remains available in the editor. Supported models include Qwen Edit, Flux 2 Klein, Nano Banana, FLUX Schnell, and Z-Image Turbo.
Choose an output shape that fits your photo or intended placement instead of being limited to square, portrait, and landscape presets.
Example: choose 4:5 for a portrait-oriented image or 3:2 for a wider photograph.
Choose an Image Editor Aspect Ratio
Open the tool to try the released workflow.
More Free Image Tools for Product Photos and Design
Explore new free image workflows for product mockups, product staging, room redesigns, flyers, invitations, and other designs. Image-editing tools accept the inputs their workflow needs; generator tools start from your subject or prompt.
Start with a task-specific tool instead of rebuilding a prompt from scratch. These tools produce image outputs; a design or presentation tool name does not imply editable slides, PDFs, or vector-file exports.
Example: upload a product image to AI Mockup Generator for a directed mockup workflow. The first-party product-photo example below appears on that tool's page and illustrates image editing, not a benchmark of the mockup preset.

Try AI Mockup Generator
Open the tool to try the released workflow.
Direct Music Videos with Timestamped Moments
AI Music Video Generator now offers Timed directions without a preview setting. Add moment-specific prompts on the timeline, then edit their timestamps and directions below the main prompt.
Guide what should happen at different points in your music video instead of describing the entire sequence in one paragraph. Save applies your moment edits; Cancel leaves the existing directions unchanged.
Example: choose a point on the timeline and write a direction for that moment, then add another moment later in the clip.
Add Music Video Directions
Open the tool to try the released workflow.
Choose a Clip Range for Video-to-Audio
Video-to-Audio now lets you set both the start and end of the clip you want to process. The selected range is preserved when you Recreate or Retry the project.
Generate audio for the relevant moment instead of processing from the beginning of your source video. Validation and credit calculation use the selected range.
Example: release validation selected seconds 40–45 of a 45-second video and produced a five-second result with audio. The first-party product demo below illustrates the Video-to-Audio workflow, not that validation recording.

Open Video-to-Audio
Open the tool to try the released workflow.
Qwen Image 2.1 for Image Creation and Editing
Qwen Image 2.1 is now available for image generation, image editing and the public API. It is the default model when starting a new AI Image Editor session.
Start an edit with Qwen Image 2.1 without finding a preview setting. Your explicit model selections remain respected.
Try a concrete edit such as changing a background or an outfit. The first-party examples below illustrate these editing workflows; they are not a Qwen Image 2.1 quality benchmark.

Try Qwen Image 2.1
Open the tool to try the released workflow.
September 2026
Qwen Edit 2K Credit Update
Qwen Edit images at 2K now cost 20 credits each, up from 15. The 640px and 1K options remain 10 credits per image.
For four 2K images, budget 80 credits; the previous rate was 60. Review your model, resolution, and credit quote before generating.
Open AI Image Editor
Review the resolution and credit quote for your next edit.
Guide Image-to-Video with Keyframes
Image-to-Video now offers Keyframes alongside Start & End Frame. Add multiple images to guide how your video moves between scenes.
Use keyframes when you need more control over the visual sequence. Switching modes preserves your keyframes, prompt and previous model selection.
For example, arrange several images in the order you want them to appear, then use Keyframes to guide the sequence. Watch the Magic Hour walkthrough below to see the workflow.
Try Image-to-Video Keyframes
Open the tool to try the released workflow.
Remove Watermarks from Videos
Video Watermark Remover is now available directly in Magic Hour. Upload a video and remove visible watermarks from the clip.
Clean up a video for your next edit without manually repairing every frame. Use it only on videos you own or have permission to edit.
Example: the first-party product demo below shows the watermark-removal workflow.

Remove a Video Watermark
Open the tool to try the released workflow.
New Model Training Retired; Existing Models Still Available
You can no longer start new custom-model training. Use saved Characters for new reference-based projects across supported tools.
Models you trained before this change remain available in AI Selfie Generator and AI Headshot Generator.
Open My Models on the former Train Model page to use, rename, or delete an existing model. For a new identity, save your reference photos as a Character.
Create a Character
Save your reference photos for supported creation workflows.
Translate Videos Through the API
AI Video Translator is now available through POST /v1/ai-video-translator. Submit a 1–30 second clip and choose from 53 target languages, then poll the existing video-project endpoint for the result.
Add video translation to your application's workflow without opening the web editor. Requests default to 480p; 720p and 1080p require a paid plan.
Example: the still below comes from Magic Hour's Spanish translation demo. Open the API page to play the translation example and see request and polling guidance.

Explore the Video Translator API
Open the tool to try the released workflow.
Longer Prompts for MiniMax H3 in the Text-to-Video API
MiniMax H3 in the Text-to-Video API now supports longer prompts, expanding beyond the previous 2,000-character limit.
Keep more scene and camera directions in one request. For example, describe a product reveal from an opening close-up to a wider scene, including the lighting and camera movement for each part.
View Text-to-Video API docs
Check current model-specific limits and build your request.
Use Five Character References in Music Videos
Music Video Generator now accepts up to five saved Characters in one prompt, up from two, on Creator, Pro, and Business.
Plan a larger cast without removing characters from your creative brief. Select saved Characters from References and mention their @references when describing your scene.
Open Music Video Generator
Bring more of your saved cast into a music video brief.
Create Music Videos on the Free Plan
AI Music Video Generator is now available at 480p on the free plan using credits. Higher resolutions remain plan-gated; free-plan access does not mean unlimited or credit-free generation.
If your balance cannot cover the full clip, the credit prompt can offer a shorter duration that your current credits cover. That lets you make a shorter music video without repeatedly hitting the same insufficient-credit message.
Example: the first-party Misty Harbor sample below illustrates the Music Video Generator workflow. Upload your audio and choose a duration to try it with your own music.

Try Music Video Generator
Open the tool to try the released workflow.
Connect Claude with Magic Hour Sign-In
Connecting Claude to Magic Hour now uses your Magic Hour sign-in and an explicit consent screen instead of asking you to paste an API key.
Select Allow to connect your account, then ask Claude to create with Magic Hour. You can manage the connection's API key in the Developer Hub.
Example: the Magic Hour MCP page includes a night-market image prompt. The first-party image below illustrates that creative workflow, not the sign-in screen.

Connect Claude to Magic Hour
Open the tool to try the released workflow.
AI Sound Effect Generator and Video Extension Workflows
AI Sound Effect Generator
Create standalone sound effects from a text prompt, choose a duration of up to 60 seconds, then preview and download the generated audio.
Open the AI Sound Effect Generator to get started.
Video Extender and Video Expander
Video Extender and Video Expander are now connected across the free-tool experience and the Magic Hour dashboard, so you can continue working with your video in the full creation flow.
Commercial Rights with Credit Packs
Credit pack purchases now include commercial-use rights in addition to watermark-free exports. You can use eligible Magic Hour outputs commercially without starting a subscription.

Generative Fill, WAN 3.0, and More Creation Tools
Generative Fill
Add, replace, or transform selected parts of an image with a prompt. Brush, lasso, and image-mask controls make it easier to target the exact area you want to change.

Try Generative Fill in Magic Hour.
WAN 3.0 in Video Tools
WAN 3.0 via Pika is now available in the Video Editor, Text to Video, and Image to Video workflows.
More Improvements
Voice tools now support more languages, the Characters page includes more creation tools, Audio Translator has a dedicated product experience, and dashboard outputs offer more follow-up actions for continuing your workflow.
Krea 2 for the AI Image Editor API
Krea 2 is now available for AI Image Editor API workflows. Developers can use it for prompt-guided image editing through the same Magic Hour API workflow.
See the Magic Hour API documentation for implementation details.
5-Credit Virtual Try-On with FLUX.2 Klein
FLUX.2 Klein is now available in Clothes Changer, with try-ons costing 5 credits at 640px or 1K. The previous Clothes Changer workflow charged a flat 25 credits.
Add a person photo and a garment photo, choose FLUX.2 Klein, and select the outfit area to replace. Signed-in Free accounts can use 640px; 1K requires Creator, Pro, or Business. FLUX try-ons support up to 1K.
FLUX.2 Klein is also available as Skin Tone Match in Head Swap and single-face Face Swap Photo, with up to 2K on paid plans.
The original September interface shows where to choose the model and review its credit cost.

Try Clothes Changer
Choose your person, clothing item, and output resolution.
Create and Edit with GPT Image 2.5 Flare
GPT Image 2.5 Flare is available in AI Image Generator and AI Image Editor on Creator, Pro, and Business. Start from a text prompt or upload an image and describe the edit.
Choose from the model’s 640px, 1K, 2K, and 4K output settings for your next image or edit.
The official Magic Hour comparison below shows GPT Image 2.5 Flare image examples alongside other models.
Try GPT Image 2.5 Flare
Choose GPT Image 2.5 Flare in the model menu.
Add Speech to Free Talking Photo and Lip Sync
You can now generate speech from text in the free AI Talking Photo and Lip Sync tools. Choose Text to Speech when adding audio, then enter your script and select a voice and language.
More Avatar Presets and Backgrounds
AI Avatar Generator now offers more preset avatars and photo backgrounds to choose from.
Other Improvements
Compare original and transformed images or videos with a before/after slider on supported outputs.
Download or recreate audio directly from free AI Voice Generator and AI Music Generator results. Recreating music restores the original prompt, lyrics, and instrumental setting.
Google Omni 1.1 for Video Generation and Editing
Google Omni 1.1 is available for text-to-video, image-to-video, video editing, and video extension. You can also guide generation with a start and end frame.
Use the frame controls to guide where a short transition begins and ends. Google Omni in the dashboard requires a Creator, Pro, or Business subscription.
For example, describe a short glassblowing shot, then choose Google Omni in the model menu. The official Magic Hour review below shows the model in use.
Try Google Omni 1.1
Open Text-to-Video with Google Omni selected.
Free MiniMax H3 Tool
The free MiniMax H3 tool is now available directly on its model page. Use a scene description to start exploring what the model can create.
Example prompt: “A ramen chef in a tiny street stall at night ladles steaming broth into a bowl, steam curling through pink and blue neon tube light, rain falling outside, no signs or lettering anywhere, slow push-in, sizzling and rain sounds”.

Watch the five-second H3 output
Try MiniMax H3
Describe a scene and try the model.
August 2026
LTX 2.5 for Video and UGC Generation
You can now select LTX 2.5 for text-to-video, image-to-video, and AI UGC ad generation.
Other Improvements
See how many generations are running above the Generate button.
Preview uploaded audio before generating by pressing play on the audio thumbnail.
Reference Uploaded Images More Easily
AI Image Editor now adds number labels to uploaded images, making it easier to refer to specific images in your prompt.
Continue Editing Video Outputs
You can now continue working from video outputs with new chaining options:
Add Subtitles opens the Auto Subtitle Generator with your video prepopulated. Free users are taken to https://magichour.ai/create/auto-subtitle-generator.
Edit Video takes you directly to the AI Video Editor.
AI Avatar Generator Is Now Live
AI Avatar Generator is now available at https://magichour.ai/create/ai-avatar-generator. Create an avatar by choosing a character and voice, generating the character in a new setting, and writing a script for the avatar to perform.
Preset characters and voices are linked, and the generator works with characters that include voices. Your avatar projects and history are managed separately from Talking Photo.
New Character Animate Free Tool

You can now try Character Animate for free directly from the Motion Control page. Switch between Character Animate and Character Replace to create new motion-controlled videos with the workflow you already know.
Try it out now: https://magichour.ai/products/motion-control
New Magic Hour MCP Setup Page
We added a public Magic Hour MCP landing and setup page with connection instructions, client-specific configuration, example prompts, supported capabilities, and links to the technical documentation and GitHub repository.
Visit https://magichour.ai/mcp to learn how to connect Magic Hour to a supported AI agent.
Translate Videos in the Create Flow
You can now translate uploaded videos directly from the Create page. Choose a target language from 54 supported languages, optionally trim the video, and receive a lip-synced translated MP4.
Videos can be up to 60 processed seconds, with usage charged at 48 credits per processed second. The source language is detected automatically.
LTX-2.3 Now Available for UGC Generation
LTX-2.3 is now available as a model option for UGC generation. All users can generate with LTX-2.3, while free users have access to 480p resolution.
Create Free AI UGC Ads with LTX 2.3

You can now create AI UGC ads for free using an image and a script. Upload an image to use as-is, add your script, and generate a video up to 3 seconds long with LTX 2.3 at 480p.
Free users can generate one UGC ad per day.
Try it out now: https://magichour.ai/products/ai-ugc-ad-generator
Add Audio References to LTX 2.3 Video Generation
You can now upload audio when using LTX 2.3 for text-to-video and image-to-video. Add a start frame to use the image-to-video flow, or generate from text without one; audio is used to guide the resulting video.
Control Voice Speed Across Audio Tools
Voice Generator, Voice Cloner, and Voice Changer now include a Voice Speed control from 0.75x to 1.30x, with 1.00x as the default and 0.01x increments. The same control is also available for text-to-speech in Talking Photo and Lip Sync.
Speed changes preserve the original pitch, so you can fine-tune delivery without making voices sound higher or lower.
Use References with End Frames in Image-to-Video
Image-to-video generation now supports combining reference images with an end frame, giving you more control over both the visual guidance and how the video finishes.
Add Characters to More Creation Modes
You can now select saved Characters in FSV, FSP, CR, BS, and TP. When adding an image in these modes, choose Add Character to open your saved Characters and use their primary image.
Save Brand Kits for Branded Image Creation
Save your logos, product photos, and visual references in a reusable Brand Kit. Bring the same brand inputs into new image briefs without uploading them each time.
For example, collect your campaign artwork and product images in a kit, then select its saved reference in a compatible image workflow.
The original Brand Kit interface lets you add visual assets, choose a primary image, and save the kit with a name.

Build a Brand Kit
Save your visual assets for your next image brief.
Added Kling 2.6 for Text-to-Video and Image-to-Video

Kling 2.6 is now available for text-to-video and image-to-video generation. It replaces Kling 2.5 as the lower-cost Kling option, and API requests for Kling 2.5 are routed to Kling 2.6.
New Photo, Face, and Skin Enhancer Tools
You can now enhance photos, faces, and skin directly from the creation dashboard. Each tool includes three transformation strengths, giving you more control over the final look.
Try it out now:
Added Seedance 2.5 for Video Generation
Seedance 2.5 is now available for text-to-video, image-to-video, and reference-to-video generation. It delivers high-quality results with improved realism, detail, and motion at 24 FPS.
The model supports 480p and 720p output. Generation costs 288 credits per second at 480p and 576 credits per second at 720p.
Create AI Music Videos
You can now turn an audio track, optional reference images, and creative direction into a complete music video. Choose story mode or lip sync mode, with support for 9:16, 16:9, and 1:1 aspect ratios.
Open Text-to-Video
Try the workflow shown in the tutorial.
Build Photorealistic AI Influencers from Traits
Create a photorealistic AI influencer with a game-style character builder instead of writing a prompt. Select traits like character type, appearance, occupation, and accessories from the left panel, then generate the result in the center preview.
Customize skin and eye colors with presets or HEX values, expand Advanced options for additional details, and keep your selections intact while generating or retrying.
Try it out now: https://magichour.ai/create/ai-influencer-generator
New AI Avatar Generator Product Page
Explore the upcoming AI Avatar Generator, designed to turn a Character or photo into a lifelike talking video with a matching voice, custom scene, and script.
The new product page covers preset and custom avatars, voice selection, scene creation, first-frame previews, script-based generation, and use cases for marketing, education, social content, and more. The tool is coming soon.
Try it out now: https://magichour.ai/products/ai-avatar-generator
Use Photo, Face, and Skin Enhancers from Their Product Pages
You can now enhance images directly from the Photo Enhancer, Face Enhancer, and Skin Enhancer product pages.
Try it out now:
Improved LTX 2.3 in Audio-to-Video

The improved LTX 2.3 model is now available in Audio-to-Video. Provide audio, a text prompt, and an aspect ratio to create video whose timing is driven by the uploaded audio.
Check it out at https://magichour.ai/create/audio-to-video
July 2026
New Free Video Colorizer
You can now colorize videos with a simple drag-and-drop workflow. The free tool supports videos up to 5 seconds at 480p, with 3 uses available per day.
Start colorizing without configuring a color strength setting—just drop in your video and get started.
Create Videos with Video Colorizer


You can now create colorized videos directly from the dashboard. Choose a video, select a resolution from 480p, 720p, or 1080p, and adjust the color strength between low, balance, and high.
Try it out now: https://magichour.ai/create/video-colorizer
Improved LTX 2.3 Talking Photo Expressive mode.
Talking Photo Expressive now uses an improved LTX 2.3 model. Provide an image, audio, and text prompt to anchor the subject’s identity and composition while the audio drives timing and lip sync.
Create Moodboards for Consistent Styles
You can now create Moodboards to capture and replicate an art style, mood, or visual vibe. Moodboards use a collage of reference images and are available alongside Characters, with presets to help you get started quickly.
Try it out now: https://magichour.ai/create/moodboards
New Free Character Replace Tool
You can now replace a character in a video using a simple two-input workflow: upload a video and an image. The tool automatically selects the character and uses the first 3 seconds of your video—no trimming required.
The free tool supports up to three generations per day at 480p. Explore the new experience at https://magichour.ai/products/character-replace
Add Image References to Google Omni Video Editor
Google Omni Video Editor now supports up to 10 additional image references alongside the primary video. Upload images through the compact Add media box, then review or remove them as compact thumbnails before generating.
The primary video remains unchanged, with inline validation for unsupported files, duration, and size limits.
Save Cloned Voices for Reuse
You can now save a cloned voice after uploading or recording audio in AI Voice Cloner. Saved voices are available in AI Voice Generator, AI Voice Changer, and in-app text-to-speech tools.
Newly saved voices appear first, followed by voices ordered by recent use, so your most relevant voices are easy to find.