Changelog

Changelog 05 rebuilt how a Motionflare video moves. This one is about what it is made of: which model writes it, what it sounds like, how long it runs, and how it reads.

A new model lineup

• Advanced: DeepSeek V4.1 Flash at 1× credits per scene, Gemini 3.8 Flash and Grok 4.6 at 2×, GPT-5.6 Terra at 3×
• Flagship: GPT-5.6 Sol at 6×, GPT-6 Astra at 12×

DeepSeek V4.1 Flash is the one to look at. It sits in the Advanced tier at 1×, the lowest price on the list, and it reads your screenshots, your logo, and your uploads directly rather than from a description of them.

It is what Auto runs on every plan now, so the same video costs half what it did last month without changing a setting. If you have been picking a model by hand to save credits, you can stop.

At the top, GPT-6 Astra is new and it is the most capable model we offer. Sol drops from 7× to 6×.

Picking a model by hand stays a Pro and Max feature.

A music library

Background music grew from six tracks to 46, across moods and tempos, licensed for commercial use. New projects start with a track chosen for the subject rather than silence, and Project Settings still takes your own upload.

90-second videos

Videos built from a website now run up to 90 seconds, up from 60. Longer pieces get a structure rather than a longer list of scenes: an opening, a middle that carries the argument, and a close that lands. There is a 15-second tier for paid social.

Captions with a style of their own

Captions used to be one bar of text, identical in every video. Each style pack now has its own caption style, drawn to match the film it sits in. Captions fill word by word with the voice and take their colour from the frame behind them, so they stay readable over a dark shot and a light one in the same video. You can also turn them off in one sentence.

Plans and pricing

• Max is $99/month for 500 credits.
• Export costs 2 credits at 1080p and 3 at 4K, down from 3 and 8.
• New free accounts get 10 credits, enough for a complete video and a revision. A free project runs 20 seconds.
• The $15 one-time export is now on the pricing page: one project, exported once at 1080p or 4K, no watermark, no subscription.

Also in this release

• Narration, music, sound effects, and captions are each yours to switch off in one sentence, at any point, without regenerating the video
• Ask for a video with no voiceover and you get one, and the narrator's voice can be changed from chat
• Over 40 new motion recipes, and a video now opens and closes on shots built for the job
• Several edits asked for in one message are charged once, not once each, and a revision that changes nothing is not charged at all
• The assistant can tell you what a change costs before it runs

Motionflare started as one thing: paste a URL, get an on-brand video. This release rebuilds how every one of those videos moves, then widens where a video can start from.

Motion Engine 2.0

Motion Engine 2.0 is a ground-up rewrite of the motion system underneath every style. Its library now holds more than 160 motion recipes: typography reveals, camera moves, product UI choreography, object movement, data animation, and the handoff from one scene to the next. For every scene Motionflare reads what the beat needs to say, chooses a compatible recipe, and adapts it to the video's style and pacing.

A style is no longer one set of moves played on repeat. Two videos can share the same visual pack without entering, emphasizing, and transitioning in the same way.

A video from a topic, with no website

Idea mode used to be the shorter path. It is now the more powerful one.

Type what the video is about and the agent does the work: it searches the web for grounding material, pulls real facts and imagery from what it finds, drafts the story, plans every scene, writes and voices the narration, and builds the finished video out of a single sentence. Web search runs on Pro and Max.

Your prompt always wins over what the agent finds, and searched material is treated as unverified external context rather than restated as your brand's own claims.

Style packs you can see first

The look is no longer something you find out about after generating. The style picker previews every pack before anything is made. There are ten packs now, three of them new: Vox, an editorial paper collage; Ink Story, hand-drawn and narrative; and Whiteboard, the explainer classic done properly. Leave it on Auto and Motionflare picks the pack that suits your subject.

Dark mode

A proper dark mode across the whole app, marketing pages included, on a warm charcoal palette rather than flat black. Pick light, dark, or system from the switcher in the navigation bar.

Bring your own images

Motionflare has always read your website. It can now read what you give it directly.

Attach up to 10 images when you create a project, or drop them into the chat afterwards. Every upload is analysed on arrival, so the assistant can see it, talk about it, and place it in a scene. Some uploads are content and appear on screen; others are style references that shape the look without ever being shown. You choose which is which.

Uploads survive a re-scrape of the source website, so refreshing the brand material never quietly drops them.

Smaller things

• Real recorded sound effects, cut to each scene's beats, in four packs: Soft UI, Punchy, Cinematic, and Tech
• New projects start with a background track chosen for them rather than silence
• Paste a finished shot list into Idea mode and it is shot line for line, with your narration kept word for word
• State the aspect ratio, length, scene count, or language in plain prose and Idea mode picks it up
• 90-second videos in Idea mode, on Pro and Max
• Change the aspect ratio after preview and regenerate from there
• 1× to 1.5× playback speed in the editor
• Project Settings split into Video and Brand tabs

A new Motionflare interface

The workspace was rebuilt around the thing you actually do: read the draft, watch the scenes, fix what is wrong. Projects, previews, and settings were rearranged so the common path is short and the rare controls stay out of the way.

A model picker

You can now choose the generation model per project. Five models across three tiers:

• MiniMax M3 — the most efficient option
• Grok 4.5 — a strong balance of speed and quality
• Gemini 3.6 Flash — fast and balanced
• GPT-5.6 Terra — advanced quality and reasoning
• GPT-5.6 Sol — maximum quality for demanding work

The picker shows the quality profile and the credit multiplier before you begin. MiniMax, Grok, and Gemini are available to everyone; Terra and Sol are Pro and Max.

Your choice stays attached to the project, so initial generation and later refinements stay consistent, and it is visible in Project Settings.

Edit on-screen text visually, for free

Small copy changes should not require rebuilding an entire scene.

The Scene Editor finds the static text rendered inside each scene and lets you edit it directly — a headline, a typo, a URL, a line of copy you want gone — then saves everything together. Visual text edits are free and preserve the existing animation, voice-over, subtitles, timing, transitions, and layout. If the new copy is significantly longer, Motionflare warns that it may affect the composition without blocking the edit.

For larger creative changes the same panel has a separate Regenerate tab.

More control after generation

Project Settings brings brand, music, and voice controls into one place: replace the project logo, manage the watermark, change background music and its volume, choose another voice, update the voice-over language, and control subtitles without starting over.

Changing the voice rebuilds only the voice assets and keeps the finished scene visuals, so you can try a different delivery without asking Motionflare to redesign the video.

Ten interface languages

The Motionflare interface is now available in ten languages, and every project handles two language choices separately: the screen-text language (English, Greek, Spanish (Latin America), German, French, Chinese Simplified, Chinese Traditional, Japanese, Korean, Hindi) and the voice-over language and accent (English US, English UK, Greek, Spanish (Latin America), German, French, Mandarin, Taiwan Mandarin, Cantonese, Japanese, Korean, Hindi).

Pick the writing system your audience reads and match it with the voice or regional accent that feels natural; Motionflare builds the video around both. Your choice is remembered, so the next time you return the product stays in the language you picked.

Subtitles that stay with the video

Many videos are watched before they are heard. Subtitles can now be turned on in the editor, so every line of narration is readable from the first frame and stays burned into the exported file.

A native 9:16 canvas

Landscape is not the natural home of every launch. Motionflare now supports a native 9:16 portrait canvas built for TikTok, Instagram Reels, YouTube Shorts, and other mobile-first placements.

It is not a crop of the landscape cut. Scenes are planned and composed for the vertical frame from the start.

Two more requests kept coming back after Changelog 01: reach, and that first quiet moment of staring at a blank workspace.

Multi-language video

A launch video should sound like it was made for the people watching it, not translated for them. And a launch does not only happen in English.

Motionflare now builds the whole video in another language. This is not subtitles and it is not an English video with a translation bolted on at the end: the on-screen scene text and the spoken voice-over are both generated natively in the language you pick. The motion, the timing, and the creative direction stay exactly as they would in English.

Today you can choose English (US), English (UK), or Greek. Greek was a deliberate first test — a different alphabet and different letterforms push the system much harder than another Latin-script language would, so getting screen text and voice clean there makes the next language mostly a matter of turning it on.

More languages are on the way.

A built-in tutorial

The first time you meet a tool that turns one URL into a finished video, the hard part is not the work. It is knowing where to begin.

So there is now a short video tutorial in the workspace sidebar, sitting next to the work instead of buried on a docs page you have to go and find. Watch it before your first project, or glance back at it when you want a reminder.

It is the smaller of the two changes here. On day one, it might be the one that matters most.

Public Beta opened and the requests came in faster than the roadmap. Three of the most repeated ones are now live.

Prompt mode

Until now Motionflare started in exactly one place: a URL. That is still the fastest path when you already have a live site, but not every idea has a landing page yet.

Prompt mode adds a second way in. Pitch a feature that ships next month, tease a concept with no page, or build the video for something that currently lives only in your head. The engine underneath is identical: it drafts the script, plans the scenes, picks a direction, and voices it. Only the input changes — intent instead of a link.

Music and voice

Music. Pick from a set of built-in tracks or upload your own. Calm and considered, quick and punchy, cinematic and wide. Whatever you choose sits under the voice-over and the motion, mixed so nothing has to fight to be heard.

Voice. Match narration to your audience: US or UK, male or female, every option with a sample you can play before you commit.

Share links

Export used to hand you a file. Now it hands you a link too.

Every export mints a public page at motionflare.ai/share. Send it in a DM, drop it in a Product Hunt comment, paste it into the group chat where the launch is happening. Whoever opens it watches right there — no download, no login, nothing to install. The MP4 is still yours to keep.

Keep it coming

Every one of these started as a message from someone using the product. You tell us where it is thin, we go fill it in.