May 2026 Highlights: the Pixio API launches · a reworked Boards & Canvas · a multimodal Prompt Optimizer · 3D rigging & animation · and a huge wave of new image, audio, avatar, and video models — with prompt examples throughout.
🔌 Introducing the Pixio API
The biggest news this month: Pixio now has a public API. Generate media from your own apps, scripts, and automations using the same models you use in the app.
- Account API keys — create and manage keys right from your account.
- A complete generation flow — list models, fetch each model's inputs, upload your own source media, start a generation, and poll for the result.
- Docs included — published API documentation so you can integrate in minutes.
Build Pixio-powered tools, batch jobs, and integrations — all on the same models and pricing you already know.
🎨 Boards & Canvas, Reworked
We shipped v1 of the reworked Boards and Canvas — a cleaner, more flexible space to lay out, compare, and iterate on your generations, with a smoother storyboard and editing experience.
✍️ A Multimodal Prompt Optimizer
The Prompt Optimizer is now multimodal — drop in an image and it factors what it sees into a sharper prompt, not just your text. Great for matching a reference look or extending an existing shot.
🖼️ New Image Models
Krea 2
A style-first foundation image model. Krea 2 Large is built for photorealism and raw aesthetics; Krea 2 Medium shines for illustration and artistic styles. It has best-in-class style transfer (feed it a moodboard of references) and a tunable creativity dial.
Try this prompt (Krea 2 Large): "Ultra-realistic portrait of an elderly fisherman at golden hour, weathered skin, salt-and-pepper beard, soft rim light, shallow depth of field, shot on 85mm, natural color grade, photojournalistic."
Recraft V4.1
A design-first model that treats typography as part of the image — so on-image text actually reads correctly — and can output native vector SVGs you can open and edit in Illustrator or Figma. Includes Pro, Utility, and text-to-vector variants, with 2K output and color-palette control for brand-consistent assets.
Try this prompt (Recraft V4.1 / V4.1 Vector): "Minimalist flat vector logo for a coffee brand named 'NORTHWIND', a stylized mountain peak inside a steaming cup, two-color palette deep navy and warm cream, clean geometric lines, centered, SVG."
HiDream O1 Image
A high-resolution model (up to 2K) that's especially strong at long, accurate text rendering and instruction-based editing and personalization. Available as the full model and a faster distilled Dev version, each with an Edit mode.
Try this prompt (HiDream O1): "A vintage travel poster for 'KYOTO', autumn maple leaves framing a pagoda at dusk, large elegant serif title text reading 'KYOTO' and a small tagline 'Japan in Fall', warm muted palette, screen-print texture."
ImagineArt 2.0
A reasoning-based model tuned for true-to-life realism as well as cinematic and artistic looks, with vivid color and precise text. A 2K reference-guided Edit preview is also available.
Try this prompt (ImagineArt 2.0): "Cinematic wide shot of a lone red sports car on a rain-soaked mountain highway at dusk, dramatic storm clouds, headlights cutting through mist, reflections on wet asphalt, anamorphic lens flare, color-graded teal and orange."
FLUX.2 Pro Outpaint
From Black Forest Labs — expand any image beyond its borders. Turn a portrait into a landscape, add headroom, or extend a scene while keeping it seamless.
Grok
New Grok Image Quality and Grok Imagine Quality Edit models for high-fidelity generation and editing.
🎵 New Audio & Voice
- Mirelo SFX 1.6 — generate sound effects from text or straight from a video (video-to-audio), plus extend and inpaint existing audio
- Gemini 3.1 TTS — natural new text-to-speech
- ElevenLabs Scribe — fast, high-accuracy transcription
Try this prompt (Mirelo SFX 1.6): "Footsteps crunching on gravel, distant thunder, and light rain building to a downpour — realistic, layered foley for a 10-second clip."
🗣️ Avatars & Face
- HeyGen Avatar 5 Digital Twin — lifelike presenter avatars from a single likeness
- Advanced Face Swap — higher-quality, more reliable face swapping
🎬 Video & Subtitles
- VEED Subtitles — automatic, styleable captions burned onto your videos
- Kling — additional models added to the lineup
✂️ Background Removal
- Ideogram Remove Background and a new RemoveBG model for clean, instant cutouts
👑 36 New Maker Mode Models (Prodia)
Maker Mode grew with a large Prodia catalog spanning image, edit, and video:
Image & Edit
- FLUX 2 — Dev, Flex, Max, and Pro (text-to-image + edit)
- Gemini 3 Pro and Gemini 3.1 Flash (text-to-image + edit)
- Recraft V4 and Recraft V4 Pro (text-to-image + native SVG)
- Seedream 4, Seedream 4.5, and Seedream 5.0 Lite (text-to-image + edit)
Video
- Seedance Pro and Seedance Pro Turbo (text-to-video + image-to-video)
- Veo and Veo Fast (text-to-video + image-to-video)
- WAN 2.7 (image-to-video + video-to-video)
🧊 3D: Rigging & Animation
Meshy can now auto-rig a humanoid character and apply animations — turn a static 3D model into a posed, moving character without leaving Pixio.
Try this: generate or upload a humanoid model, auto-rig it, then apply a walk or wave animation to export a ready-to-use animated character.
📱 Mobile
The mobile app was upgraded to Expo SDK 56, keeping iOS and Android current with the latest platform improvements.
That's May — thanks for building with Pixio. We'd love to see what you make with the new API.