// AI TOOLS · GOOGLE MIXBOARD

Google Mixboard

A prompt-driven, Pinterest-style AI concepting board from Google Labs, powered by the Nano Banana image model, retired on September 28, 2026.

Last verified · 2026-10-23 · by Moe Ameen

Published 2026-09-02 · Updated

Short answer

Google Mixboard was an experimental AI concepting board from Google Labs, launched as a U.S. public beta on September 23, 2025 and shut down on September 28, 2026. You started from a prompt or a sample board, then refined generated images with Nano Banana using plain-language edits. Google said all boards, images, uploads, and prompts would be deleted after the shutdown.

Key takeaways

  • Google Labs launched Mixboard as a U.S. public beta on September 23, 2025.
  • It used the Nano Banana image model for natural-language edits, "regenerate," and "more like this."
  • Google shut Mixboard down on September 28, 2026.
  • Google said boards, generated images, uploads, and prompt history would be permanently deleted.
  • For lasting work, keep reference images somewhere durable, such as Kompozy.

What was Google Mixboard?

Google Mixboard is an experimental AI concepting board from Google Labs. Think of it as a Pinterest-style workspace where, instead of pinning images that already exist, you describe what you want and the tool generates the visuals. You start from a text prompt or a pre-populated example board, and Mixboard fills a canvas with generated images to react to.

From there you refine: mix in your own uploads, add AI-generated text blocks, and edit with natural language — "combine these two," "make it warmer" — plus one-click "regenerate" and "more like this" to explore variations fast.

The image generation and editing run on Nano Banana, Google's Gemini image model, which is what lets you make complex edits and merges by typing rather than by hand. The output is a downloadable board: you can export it as a PNG or as a presentation, and share it via a link. Mixboard launched as a U.S.-only public beta on September 23, 2025.

One critical caveat: Google shut Mixboard down on September 28, 2026. The Labs page now says the experiment has ended, and Google said all boards, generated images, uploaded assets, and prompt history would be permanently deleted from its servers after the shutdown.

For content work, keep Mixboard's role straight: it is an ideation and visual-exploration tool. It helps you decide what a look, palette, or concept should be. It does not caption or reframe video, hold a brand voice across a set of posts, fan one idea into a carousel, blog, and newsletter, or schedule and publish anywhere.

What could you make with Google Mixboard?

  • Prompt-generated mood boards and visual-direction concepts for a brand, product, event, or campaign
  • Style, color-palette, and composition explorations you iterate with natural-language edits
  • Merged and remixed image concepts (combine your reference photo with an AI-generated scene)
  • AI-generated text blocks — taglines, theme notes, naming ideas — pinned alongside the imagery
  • Downloadable boards exported as a PNG or a presentation to hand off to a designer or producer

How does Kompozy turn a Mixboard-style concept into content?

Mixboard is where you figure out what something should look like; Kompozy is where that look becomes finished, published content. The handoff is a reference image. Once a board lands on the palette, composition, or product concept you want, export that image and bring it into Kompozy as a brand reference — that single asset is enough to drive a whole production run, and because Kompozy stores generated media durably, none of it evaporates on a shutdown date.

Concretely: feed a Mixboard concept image into Kompozy's face-locked image formats and it renders Persona Photos and Persona Infographics that keep your influencer's face consistent while matching the direction you set — Kompozy uses Gemini face-lock for exactly this. Take the same look into a brand-exact Carousel rendered through HyperFrames, or into quote cards, a blog article, and a newsletter, all governed by one Persona Brief so the batch reads as one creator rather than a pile of one-off generations. Then Autopilot schedules and publishes the finished set across the eight social platforms plus blog and email behind a per-post review. Mixboard narrows a hundred visual ideas down to one direction; Kompozy turns that direction into a week of on-brand posts and ships them. (For another Google image workspace, see Google Pics.)

  1. In Mixboard, prompt and iterate until a board captures the exact look — palette, style, composition — you want, then export the winning concept as a PNG.
  2. Bring that concept image into Kompozy as a brand reference (add it to your workspace assets or a Persona).
  3. Pick the finished formats: face-locked Persona Photos, a brand-exact carousel, quote cards, a blog, and a newsletter that all match the direction.
  4. Let Kompozy generate each in your voice via your Persona Brief so the whole set stays on-brand, adding your AI-disclosure line at the per-post review gate.
  5. Schedule and publish the batch across the eight social platforms plus blog and email from one queue on Autopilot.

What does this look like in Kompozy?

This persona photo came out of the Kompozy workspace as is. Its subject has nothing to do with Google Mixboard, so look at it for the format, not the topic.

Persona photo post generated in Kompozy
Real output · Kompozy Persona photo format · from the Kompozy workspace

Who is this for, and who should skip it?

Best for:

  • Readers checking what Mixboard did and what happened to their boards
  • Creators looking for a replacement mood-board workflow
  • Teams that kept exported boards and want to use them as references

Not ideal for:

  • Anyone hoping to sign up today, since Mixboard has shut down
  • Anyone trying to recover deleted boards
  • Workflows that need finished, published posts rather than concepts

Frequently asked questions

What is Google Mixboard?

Mixboard is an experimental, prompt-driven AI concepting board from Google Labs — a Pinterest-style workspace where you start from a text prompt or template, generate a set of image blocks, mix in your own uploads and AI text blocks, and refine with natural-language edits. Its image generation and editing run on Nano Banana, Google's Gemini image model.

Is Google Mixboard being shut down?

Yes. Google shut Mixboard down on September 28, 2026. Google said all boards, generated images, uploaded assets, and prompt history would be permanently deleted after the shutdown, and shared links now show a shutdown notice.

Can I still use Mixboard?

No. Mixboard was an experimental Google Labs public beta, and it ended on September 28, 2026. For a repeatable visual workflow, use a tool that stores your images durably.

How do I turn a Mixboard concept into finished, publishable content?

Export the concept image from Mixboard, then bring it into Kompozy as a brand reference. Kompozy renders face-locked Persona Photos, brand-exact carousels, quote cards, a blog, and a newsletter that match the look, holds them to one Persona Brief, stores them durably, and schedules and publishes across nine destinations.

How do we know this?

We checked every figure and product claim on this page against the sources listed below on 2026-10-23, and removed or corrected anything they did not support. The page was written before the shutdown and now reflects that Mixboard has ended. Kompozy plan and platform details come from kompozy.io/pricing and kompozy.io.

Sources

  1. Google: Introducing Mixboard (read 2026-10-23)
  2. Google Labs Help: Mixboard shutdown (read 2026-10-23)
  3. Google Labs: Mixboard (read 2026-10-23)
  4. Kompozy: Home (read 2026-10-23)

Related tools

  • Google Pics — Google's prompt-first AI image generator and editor inside Google Workspace — describe a design in words instead of building it from a template, powered by the Nano Banana model.
  • Midjourney — The text-to-image generator known for aesthetic quality and art direction — now also building a separate medical-imaging division.
  • Google Veo 3 — Google DeepMind's video model that was the first to generate synchronized native audio — dialogue, sound effects, and music — inside the same pass as the video, with lip sync.
  • Cartesian by Formas — Formas' AI 3D-modeling tool that turns words, sketches, and dragged-in references into precise, editable models — native NURBS solids aimed at architecture and product design, not polygon-mesh blobs.
  • ChatGPT Images 2.5 — OpenAI's new state-of-the-art image model — faster generation, in-chat Sketch and Templates, and two API models (GPT-Image-2.5 Flare and Sunburst).

← All AI tools · Get started →