What's new on AICreate

What changed: A real vector editor in your browser. Open any SVG file or start from scratch. Draw with the pen, add rectangles, ellipses, lines, polygons, and text. Select shapes and combine them with boolean operations — union, subtract, intersect, exclude. Double-click any path to enter node-edit mode and drag individual anchors or bezier handles to reshape the curve, alt-click to toggle smooth/corner, click the stroke to add a new anchor. Resize with corner/edge handles, rotate with the rotate handle (Shift snaps to 15°). Choose a background color or work transparent. Reorder, lock, and hide layers in the side panel. Edit fill, stroke, dasharray, opacity, and exact x/y/w/h for any selection. Snap to grid for clean alignment. Export as a true SVG (every shape is a real path) or a sharp PNG at 2× scale. Built on Paper.js. 100% browser-based — no upload, no watermark, no signup.

Why: Drawing Tool is great for sketching. SVG Editor is the next step — open existing SVGs, edit individual nodes and bezier handles, run boolean operations, manage layers — the primitives you actually need for icons, logos, and clean shape work. Built on the most mature open-source vector engine in the browser.

What changed: A clean vector drawing tool with an infinite canvas. Native SVG document — every stroke, shape, and text element is a real path, never a raster stamp. Six pen presets (Pen, Fineliner, Marker, Ink, Calligraphy, Highlighter), shapes, real-transparency eraser, and unlimited pan and zoom. Export a true scalable SVG you can open in Illustrator, Figma, or Inkscape and resize to any size with no quality loss. Powered by ThorVG WASM — runs entirely in your browser, no signup, no upload, no watermark.

Why: The original drawing tool is built for sketching with raster brushes. This one is the clean-vector counterpart for logos, icons, diagrams, and anything you need to scale — built around an SVG document instead of pixels.

What changed: Pro-grade browser drawing app. Pressure-aware brush with stabilizer (so jittery hands still draw smooth curves), four brush presets (Pen, Pencil, Marker, Ink), shapes (line, rectangle, ellipse — filled or stroked), a text tool with font and size picker, and 4-way symmetry mode for mandalas. Drag-and-drop or paste any image to start from a photo. Custom colors with recent and saved palettes that persist between visits, and a screen-wide color picker on supported browsers. Choose any background color or work on a transparent canvas — the eraser produces real transparency, so you can export logos and stickers with a clean alpha channel. Download as a PNG or as a true SVG — every brush stroke, shape, and text becomes a real vector path, eraser strokes become an SVG mask, so you can scale your drawing to any size without losing quality. Pan, zoom, and a 25-step undo. When you're done, hand the canvas straight to AI Restyle (i2i), Image Filters, Object Remover, Background Remover, Upscaler, Vectorizer, OCR, or Crop — opens in a new tab so you don't lose your work.

Why: A free in-browser drawing app that's actually usable for real work — pressure stylus, smooth-curve stabilizer, shape tools, text, symmetry, and a one-click hand-off into the rest of the AI toolkit.

What changed: Paste any JSON and see it as a pan-and-zoom mindmap, a collapsible tree, or formatted text. The mindmap collapses anything past depth 2 by default so even huge configs stay readable — click any node to expand. Tolerant parser handles trailing commas and // comments. Search by key or value, copy any node's path, export the mindmap as SVG or PNG. Files never leave your browser.

Why: Most JSON viewers throw a wall of text at you. We already render mindmaps for ideas and markdown — the same engine works beautifully for JSON shape, and seeing the tree is much faster than reading 5,000 lines of indentation.

What changed: Side-by-side JSON comparison. Paste the original on the left and the updated version on the right, then see every added / removed / changed leaf color-coded in a single tree. Toggle 'hide equal subtrees' to focus only on what's different. Export the delta as an RFC 6902 JSON Patch — apply it programmatically with any standard library. 100% browser-side.

Why: Diffing two JSON blobs in your editor means scrolling. A structural diff with one-click JSON Patch export turns a 30-second comparison into a 2-second one — and the patch is the canonical way to ship the delta to a service or teammate.

What changed: Three ways to point at what you want gone: paint with a brush, pick a class ("People", "Cars", "Hair", "Sky", and 30+ others) for an automatic mask, or just click on the thing and let SAM 3 trace it. Then AI inpainting fills in what should be behind, plausibly enough that it usually doesn't read as edited. Each Remove pass runs on the previous result so you can iterate one spot at a time, and the whole thing runs inside the browser tab so your photos stay on your device.

Why: Until now we had a Background Remover for the whole subject and nothing for the small annoyances. Wanting to delete one specific thing without uploading to a stranger's server is the most-asked-for image tool in this category; this is that.

What changed: The AI Story Generator now exports a real PDF book: cover page with the protagonist portrait, scenario header, and one page per beat with the generated visual on top and the narration / camera / sound notes beneath. The 'Open in Book Trailer Maker' button now hands the same PDF straight to the Book Trailer Maker — no more empty trailer page when you click through.

Why: Clicking 'Open Book Trailer Maker' from the story used to land you on an empty page because the trailer needs a PDF; now we build the book for you and pass it through, and you get a downloadable keepsake too.

What changed: Type a one-line prompt and get 5 distinct story ideas — each with a logline, protagonist, conflict, and twist. Click any idea to expand it into a 15-second video scenario: 3 timed beats with visual / camera / sound / narration / caption per beat, and shared persona+style locks so all 3 frames look like the same film. Generate the protagonist portrait and each beat's visual right inside the page (no tab-hopping), play the narration in-place, and export everything as a plain .md or as a .zip with the images bundled.

Why: Suggested by a user who wanted 'inspiration for potential stories' — five short ideas to choose from beats one long story you didn't ask for, and the 15s scenario step turns the idea into something you can actually film without leaving the page.

What changed: Drop a .md file or paste text and watch it turn into an interactive mind map — headings build the spine, links between sections become cross-edges. Click any node to read or edit the section, with optional one-click AI rewrite, summarize, and explain. Export as PNG, SVG, or JSON.

Why: Long markdown docs (READMEs, notes, specs) get unreadable in plain scroll; seeing the whole structure at a glance and jumping straight to a section makes them usable again.

What changed: Each cell now has an opacity slider and a 12-mode blend menu (multiply, screen, overlay, difference…) that match exactly between live preview and the exported video or image. You can also zoom a layer below the cell so it floats inside the frame instead of always covering it.

Why: Composite looks (double exposure, color grades, subtle overlays) and shrunken floating photos previously needed a separate editor — now the entire layered look stays in one tool.

What changed: Tap any cell and pick multiple photos or videos at once — the collage fills cells in order from where you tapped. In Freeform / Picture-in-Picture layouts, extra files become new overlay layers automatically.

Why: Building a 9-grid or layered PiP one file at a time was tedious; selecting everything in a single picker matches how iOS and Android Photos already work.

What changed: Clip any layer to a circle, heart, star, or rounded square — drag/resize keeps working past the mask edge. New selected-layer panel groups shape, z-order, and remove together; the background layer is now locked at the bottom so overlays can't disappear behind it.

Why: Heart-shaped reaction cams and circular avatars were the most-asked PiP feature, and accidentally pushing an overlay behind a full-frame background made it look like the layer had been deleted.

What changed: Pick the new Freeform layout to drag, resize, and stack overlays anywhere on the canvas — no fixed grid. A dedicated /picture-in-picture/ page mounts the same engine pre-configured with a background + corner overlay.

Why: Reaction videos, gameplay PiP, and webcam-over-tutorial layouts didn't fit the old grid presets. Freeform places each video exactly where you want it without forcing a 2x2 or side-by-side template.

What changed: Drop a PDF and get an animated trailer with page-flip, Ken Burns, 3D stack, or cover-zoom effects. 9:16 / 1:1 / 16:9 export for TikTok, Reels, and YouTube.

Why: Authors, BookTokers, and brands keep asking for a way to turn a book or lookbook PDF into a scroll-stopping video without uploading the file or paying for Animaker. This runs entirely in the browser, no watermark.

What changed: Drag any cell to reposition the photo or video inside its frame, and use the in-cell slider to zoom in up to 4× — works for both images and videos, in the same collage.

Why: Center-cropped covers always cut off heads, faces, or product edges. Now you can show the right part of every clip without re-cropping the source files first.

What changed: After removing the background from your video, swap it for a clean white or black backdrop with one click — no need to fish around in the color picker for the two most common solid backgrounds.

Why: Greenscreen and bluescreen are useful, but most people just want their subject on plain white (product, talking-head) or plain black (cinematic). Now those are right next to the others.

What changed: After AI removes the background, optionally swap in white, black, a custom color, a two-stop gradient, or your own image. Applies to the preview, single download, and the full-batch ZIP export.

Why: Most people don't actually want a transparent PNG — they want their subject on a clean white card, on brand color, or on a backdrop. Now that's one click instead of opening another editor.

What changed: The /tools/ page now adapts to the selected category: hero, headline, accent color, and a rotating spotlight of featured tools all retune to what you're browsing. Tool cards switched to crisp Lucide icons.

Why: The old static featured card showed the same tool no matter what you were looking for. The new hero teases real tools from the active category so it's easier to discover what's relevant without scrolling.

What changed: Collage video exports now offer a WebM and an MP4 download side-by-side. Animated stills with Ken Burns play back reliably everywhere.

Why: Some browsers' MediaRecorder produces a fragmented MP4 that certain players reject, especially when the collage is just stills with no audio. Recording a WebM in parallel guarantees a file that always plays — and you still get the MP4 for sharing.

What changed: The trimmer now shows an audio waveform over the timeline, detects silent gaps and scene cuts with one click, fits a long video into a target duration, and lets you undo/redo any edit. Select a segment to resize, move, remove, or restore it.

Why: Manually hunting for silent pauses or scene changes in a long recording is tedious; auto-detect + a proper selection and history model turns a 20-minute edit into a 30-second one.

What changed: Trim the ends, cut a middle section, or splice several keepers into one clip. Split the timeline anywhere, toggle each segment keep or remove, drag cut markers to fine-tune, and export — no watermark.

Why: Trimming and cutting is the single most common thing people want to do to a video, and most free browser tools either limit you to the ends only or quietly watermark the output.

What changed: Record your screen, a single window, or a browser tab — with optional microphone and system audio — and download as WebM or MP4, right from the browser.

Why: Every other web screen recorder either slaps a watermark on your video, caps you at a few minutes behind a signup, or uploads to their servers first. This one does none of that — the browser's own screen-capture API does all the work, and the file never leaves your device.

What changed: Extract the audio from any video — MP4, MOV, WebM, MKV, AVI — and download as MP3 or WAV. Trim to a clip, pick a bitrate, all in your browser.

Why: Every other video-to-MP3 site uploads your file to someone else's server. This one doesn't touch the network — the whole conversion happens in the tab you're already on, so it's faster and nothing leaves your device.

What changed: On phones, the curve editor now sits transparently over the photo so you can drag control points with your finger and watch the preview change underneath. HDR sliders overlay the preview the same way.

Why: Before, the curve editor was below the preview and only responded to mouse clicks — on mobile you couldn't drag points and couldn't see the photo while adjusting. Both are fixed.

What changed: HDR Filter now shows the original next to the processed result, and video previews auto-play in a loop so you can see every curves and tone-map tweak in real time.

Why: Judging HDR and curves adjustments against the original is the whole point — having to click a Preview button and only seeing the processed side made it hard to tell what actually changed.

What changed: Convert SDR photos and videos to HDR, or fine-tune tone and color with Photoshop-style RGB curves. Combine both for color-graded HDR. Export JPG, OpenEXR, AVIF, or HDR video.

Why: HDR displays are everywhere now — iPhones, Macs, Android — but there's no free browser tool to convert existing SDR media or do proper curves grading. Now there is.

What changed: Hand a generated image straight from the AI Image Generator into Image Filters or the Product Ad Video tool with one click — and any tool can now be deep-linked from an external URL (e.g. /image-filters/?input=...&preset=vintage), which means AI agents can drive them like an API.

Why: Most real workflows are more than one tool. Forcing people to download, re-upload, and rename a file between steps is friction. Making the tools deep-linkable also means an LLM can compose URLs to run them — no API key or signup needed.

What changed: Blur, pixelate, black-bar, or emoji-mask faces in photos, videos, and webcam streams — with a shape that follows the actual face contour and emoji stickers that tilt with the head.

Why: Online face-blur tools either upload your photo to a server or draw a crude rectangle. Privacy shouldn't require trusting a server, and the fix shouldn't look like a redaction bar from a 90s news broadcast.

What changed: Generate smooth 2×, 4×, or 8× slow-motion from any short clip. AI optical-flow frame interpolation (RIFE) runs in your browser on WebGPU — it synthesizes real intermediate frames instead of just duplicating existing ones.

Why: The Video Speed Changer's slow modes only duplicate frames, which looks choppy on motion. A real slow-motion tool has to invent the missing frames, and doing it client-side means your video never leaves your device.

What changed: Type a prompt and get a polished multi-scene motion-graphics video — headlines, per-scene imagery, transitions, music, and effects. 12 templates, three aspect ratios, exports as MP4, WebM, GIF, or self-contained HTML.

Why: The Banner Generator makes a single animated slide; Studio makes a real multi-scene video with transitions and music, rendered on your own GPU so the clip never leaves your device.

What changed: The Video Converter now produces animated GIFs directly in your browser with a per-frame optimized palette, no server round-trip or upload.

Why: Server-side GIF conversion meant uploading the whole video and fighting nginx size caps; keeping it local is faster, private, and the optimized palette keeps file sizes sane.

What changed: Media Collage now animates still images with a slow zoom and pan, turning a grid of photos into a video — each cell gets its own direction so the motion doesn't feel uniform.

Why: A static photo collage is a single dead frame. Ken Burns is the classic trick for giving stills life without stock motion graphics, and now it's one click.

What changed: Font Editor adds Weight Offset (true bolder/thinner on any static font), Arc Bend (per-glyph curve), Stencil (horizontal bridge gaps), and Perlin Wobble (organic hand-drawn waver) to the Derive New Font panel.

Why: The existing effects leaned toward rough/glitchy looks. These four cover the gaps: proper weight control without a variable axis, stylized curvy type, stencil-plate lettering, and natural-looking imperfection instead of twitchy jitter.

What changed: AI text and vision tools now respond instantly via our servers while the in-browser model quietly downloads in the background — no more waiting for a 3.7 GB download before your first click goes through.

Why: The first time you opened a writer, memo, or caption tool, the Generate button blocked for several minutes while weights downloaded. Now you get your result in seconds, and later calls switch to the local model automatically once it's ready.

What changed: The Font Editor now lets you edit the font's name, family, designer, copyright, license, version and more before downloading — your changes are written into the TTF's name table.

Why: Sharing a derivative font without updating its metadata means it shows up in font menus under the original name and credits. Editing the name table closes that loop.

What changed: Beyond editing glyphs, the Font Editor now generates entire derivative fonts with one click — slant, jitter, pixelate, wave, rotate, round, extrude, scramble — plus every Unicode block browsable in a sidebar and live controls for variable-font axes.

Why: Designers want a new typeface, not a tweaked letter. Turning existing fonts into fresh variants in one click beats hunting for a font you'll pay for and still have to customize.

What changed: A browser-only photo editor with 20+ one-click looks — vintage, black & white, tilt-shift, duotone — plus adjustable LUTs.

Why: Most people just want a good-looking photo in 10 seconds, not a Photoshop session. Pick a filter, download, done.

What changed: Load any Google Font directly into the editor, then tweak individual glyphs, weight, or spacing before exporting.

Why: Everyone starts from an existing font — uploading a local TTF was a pointless hurdle. Now you can go straight from "I like this font" to "I want to tweak this one letter".

What changed: Overlay animated text on any video — custom fonts, colors, animations, and timing — right in your browser.

Why: TikTok-style captions and lower thirds shouldn't require a $20/month app or a server upload. This handles the 90% case for free.

What changed: HEIC (iPhone) and AVIF files are now accepted as input and decoded on your device.

Why: Every iPhone user hits this: you drag in a photo and nothing happens because it's HEIC. Now it just works — no separate converter app.

What changed: After the image generator returns multiple variations, one button packs them all into a single download.

Why: Nobody wants to right-click-save four times in a row. One click, one file, move on.

What changed: HTML, business plan, mindmap, sticker, icon, and banner generators now show clickable example prompts.

Why: A blank input is the #1 reason people bounce. Seeing a good example takes you from "I don't know what to type" to a result in seconds.

What changed: Paraphrase, summarize, email writer, ad copy, meta descriptions, grammar checker, and friends all use a local AI model — and each ships with example prompts.

Why: Your drafts shouldn't ride to a server and back. Local means private, fast, and free, with no daily cap.

What changed: The image generator gained new one-click styles including Cartoon plus seven photorealistic looks with studio-quality camera, lens, and lighting tuned in automatically.

Why: Writing a prompt that actually looks good is its own skill. Presets shortcut that — pick "Cartoon" or "Cinematic" and the rest is handled.