Wildfield is an HTML ad builder with AI image generation, brand design systems, and one-click multi-format export. Compose scenes, generate visuals, render video, and ship on-brand creative — by hand in the studio, or fully automated by API.
Free to start · Pay per generation, or add a plan from $20/mo · Hand a brief to an agent over MCP · AI imaging, physics motion & sound · Export PNG, video & Figma-ready HTML
AI visuals that stay perfectly on brand.
Start free →From brand setup to final render — Wildfield folds design systems, AI generation, scene composition, and export into a single fast workspace.
Store colours, font stacks, logos, CTA styling, typography weights, transforms and letter-spacing once. Every ad inherits your brand automatically — no drift.
Colour vars · 6 font roles · logo positioningSquare, Portrait, Vertical, Landscape and Wide — one project, every placement. Type scale and padding adapt per format, and per-platform safe-zone guides keep copy clear of the UI.
5 formats · safe zones · auto-responsive typeBuild ads as a sequence of scenes — headline, subhead, H3, CTA, background and overlays per scene. Sequence them into a timeline for static sets or video.
Per-scene styling & timingText-to-image across GPT Image 1.5, Flux, Imagen 4, Nano Banana Pro and Ideogram — with camera, lens, lighting, shot-size and mood direction built in.
7+ models · art direction controlsColour, gradient, image or folder on a scene. Background video is a project-level track: one clip under the whole timeline, so cuts stay clean in preview and in the cloud render. Stream a tab onto the canvas, then add overlays, SVG, glyphs, crop and flip.
Project-level video · live stream capture · layersShip as high-res PNG, cloud-rendered MP4 video, or Figma-ready HTML. Server-side rendering means pixel-perfect output every time, at any scale.
PNG · MP4 · Figma HTMLEight forces push copies of a glyph or SVG around the frame: wind, turbulence, gravity, swirl, orbit, wave, float, spring. Recipes, Perfect loop, trails and a Guide (form a ring, follow a path) sit on top. Linked scenes hold then morph. Preview matches the export.
Forces · recipes · linked morphDrop an mp3 or wav onto the canvas and the timeline grows a live waveform. Score the whole ad or drop audio onto a single scene, drag the selection to line it up, then crop and fade.
Timeline & per-scene · crop & fades · scene-cut guidesA design system is the source of truth for a client. Define the palette, type and logo, and every scene, format and AI generation inherits it — so creative scales without ever going off-brand.
Generate, extend, upscale, retouch and vectorise — without ever leaving the canvas. Powered by fal.ai.
Describe it, direct it with camera, lens, lighting & mood, and pick from 7+ models.
Expand any image to fill a new format — top, bottom, left or right — seamlessly.
Aura-SR, creative and clarity upscalers push images to crisp, print-grade detail.
Brush out distractions or paint in new elements directly in Imaging mode.
Switch models per generation — fast drafts on Flux Dev, hero shots on GPT Image 1.5 or Imagen 4, typographic layouts on Ideogram.
Need a clean icon, badge or shape instead of a photo? Generate crisp, infinitely-scalable SVG illustrations with Recraft V4.1 on fal.ai — drop them straight onto the canvas as layers.
SVG generation · editable layersInput → Layout → Canvas → Output. Four focused modes, one render path.
Wildfield keeps the workflow linear and fast — write copy, arrange the layout, art-direct the canvas, then export. What you see is exactly what renders, because preview and output share the same engine.
A vector layer becomes a field. Wind carries every copy the same way. Turbulence makes it jitter. Gravity drops it. Swirl and orbit spin it. Wave and float roll it. Spring pulls it home. Recipes write those forces for you. Perfect loop locks the last frame to the first. LINK holds each scene's look until the keyframe, then morphs into the next.
Drop a token into any vector text layer and it animates live over the scene — a number that counts, words that cycle, a random digit that stays put. No keyframes. What you preview is exactly what renders, and it carries straight across linked scenes.
{0-100} number counts up or down
{cat, dog, bird} words take turns
{random} stable digit
{type:SALE} types on
{dogs,cats:s0.5} stagger each letter
Mix tokens with plain text in one line — plain text is left untouched. Read the syntax →
Drop audio straight onto the canvas and Wildfield drops you into Sound mode. See the waveform, crop and preview the section you want, set a threshold, then analyse the transients — Wildfield generates scenes that land on the beats, and you can nudge any of them. Layer a voice-over on top and mix.
Newest module — rolling out now. Music playback & the per-scene audio library are live today.
Static, motion or hand-off — Wildfield renders server-side for pixel-perfect, repeatable output.
Server-rendered screenshots at full format resolution — ready for every social placement and display network.
Sequence scenes into motion ads and render MP4 in the cloud via a distributed HyperFrames Lambda pipeline — with audio.
Export clean HTML compositions to hand straight to design — for review, tweaks or pixel-level polish in Figma.
Every action in the studio is an endpoint. Drive Wildfield from your own tools, pipelines, or an AI agent — with a single API key.
# Generate an on-brand AI image curl -X POST https://wildfield.io/api/genai/generate \ -H "X-API-Key: adk_•••••••" \ -H "Content-Type: application/json" \ -d '{ "prompt": "luxury watch on marble, studio light", "model": "gpt-image-1.5", "width": 1080, "height": 1080 }' # → { "ok": true, "path": "https://…blob/img.png" }
curl -X POST https://wildfield.io/api/projects/save \ -H "X-API-Key: adk_•••••••" \ -d '{ "filename": "june-campaign", "project": { "version": 2, "designSystemId": "acme", "activeFormat": "square", "scenes": [ … ] } }'
Hand a brief to Claude, Cursor, or Grok. The agent sets up the design system, scenes, imagery, physics, and a render. One API key.
create_design_systemcreate_projectgenerate_imageget_motion_skillgenerate_vectorcapture_urlget_credit_balancerender_video
Auth via X-API-Key header · keys start with adk_ · CORS-enabled.
Sign in with Google or email. Admins assign design systems to users — everyone sees only the brands they own. Clean separation, zero confusion.
Work is saved continuously to the cloud and your browser. Pick up any project from any machine — projects and design systems persist server-side.
One HTML composition for preview and export. No never-ending spinners. Go from a brief to a rendered ad in minutes.
Built for ads, not retrofitted into one. Figma designs interfaces and Canva makes template graphics — Wildfield turns a brand into on-brand, animated ads at scale. Here's where each one fits.
Figma Motion is a deep, keyframe-level tool — designers tune every easing curve by hand, and it's superb at that; Canva is great for fast template graphics. Wildfield runs on the opposite axis: its agent produces smooth, well-eased, physics-driven motion on autopilot — no keyframes or curves to fiddle with — brand-applied and rendered to video in every ad format, with sound. If your job is shipping on-brand ads at scale, that's the gap Wildfield fills. Start building free →
Building, editing and PNG & HTML export are free on every plan. AI generations and video renders draw from a prepaid credit balance you top up from $5, on any plan including Free. A subscription buys design-system capacity, and drops credits into that same balance every month.
Credits pay for the AI on every plan, Free included. Plan credits land in the same balance, so there is one number to watch and one set of prices.
Prices in AUD and include GST. Plans are per account, billed monthly, and buy design-system capacity only. Every generation shows its cost up front and draws from your credit balance, so you can track every cent in your billing panel.
Set up your brand, generate the visuals, and export PNG, video and HTML — all in one place. Or wire it into your stack with the API.