About Flato

Core Definition

Flato is a persistent structured canvas project for AI agents and human editors.

A Flato project can contain pages, blocks, text, images, video, shapes, layouts, styles, themes, animation definitions, interaction behaviors, and project-level state. This structure lets AI agents inspect an existing design, make targeted edits, and continue work without flattening the result into a disconnected file.

Flato is presentation-centered, but not presentation-only. Presentations are the primary scenario because they combine narrative, layout, text, visual hierarchy, brand consistency, and repeated iteration. Flato extends that same canvas model into interactive slides, web-style visual experiences, brand assets, motion graphics, and video-oriented visual content.

Canvas as Harness

Flato treats the canvas as a harness for AI design.

The canvas gives AI agents structure: pages, blocks, layouts, styles, media, motion, interactions, and project state. Instead of generating from an empty prompt, agents work inside a constrained visual workspace, which helps them make more targeted, editable, and design-aware decisions.

Canvas as AI-Human Shared Workspace

The canvas is the shared workspace between AI agents and human editors.

AI agents can inspect and update the project structure. Humans can review, adjust, export, and continue the work manually. Both sides operate on the same persistent canvas, so collaboration happens through the design itself rather than through disconnected files or regenerated outputs.

Product Forms

Flato has two main product forms: Flato MCP and Flato SaaS.

Flato MCP is the agent-facing runtime and a core part of Flato's value. It connects MCP-compatible AI agents to the Flato canvas runtime. Through MCP, agents can create new projects, take over existing editor or share links, inspect the current canvas, edit specific pages and blocks, and continue improving the same live project over multiple rounds.

Flato SaaS is the browser-based editor. It lets human users open the same projects, review AI changes, make manual edits, adjust layouts, update content, export deliverables, share links, and continue collaborating with AI.

Both product forms operate on the same project state. The agent and the human editor are not working on separate copies of the design.

How Flato Works

In a typical Flato workflow:

• A user or AI agent starts from a brief, an existing project, an editor URL, or a share link.
• Flato creates or opens a persistent editable canvas project.
• The AI agent reads the current canvas state, including pages, blocks, layout, styles, media, and project structure.
• The agent creates or edits specific pages and blocks instead of regenerating the whole design blindly.
• Human editors can review, adjust, resize, crop, rewrite, restyle, and export from the same project.
• The project can continue across multiple rounds of AI and human editing.

What Flato Can Create

Flato supports presentation-centered visual work that needs to stay editable after generation.

Flato can be used for:

• Presentations and slide decks.
• Interactive slides and visual storytelling.
• Interactive visual pages and web-style experiences.
• Social media visuals.
• Brand assets and marketing visuals.
• Posters and infographics.
• Motion graphics and animated content.
• Video-oriented visual content.

These formats stay connected to the same editable canvas model rather than becoming disconnected outputs.

Agent Capabilities

Through Flato MCP, an AI agent can:

• Create a new editable Flato project.
• Take over an existing editor URL.
• Work from a supported share link.
• Inspect the current canvas context.
• Edit pages, blocks, text, media, layout, and styles.
• Add or adjust charts, animations, and interactions where appropriate.
• Prepare or export deliverables from the same canvas.
• Continue improving the same project over multiple turns.

Human Editing Capabilities

In the browser editor, human users can:

• Review AI-generated work.
• Edit text and layout manually.
• Move, resize, crop, and arrange content.
• Insert text, images, video, and other visual elements.
• Adjust canvas and style settings.
• Export and share the project.
• Continue collaborating with AI on the same canvas.

Built-In Creative and Rendering Infrastructure

Flato includes creative resources and rendering capabilities directly inside the runtime:

• Fonts and multilingual typography resources, including curated CJK support.
• Stock image search across providers such as Unsplash, Pexels, and Google.
• Image generation and editing through supported models such as GPT-Image-2, Gemini, Seedream, and Ideogram.
• Editable charts and data visualizations.
• Animation powered by GSAP.
• Interactive behaviors for slides and visual experiences.
• PNG image export.
• PDF export.
• MP4/MPEG-4 video export.
• Image-based PPTX export.
• Online shareable presentation outputs.

Current PPTX export is image-based rather than fully editable PowerPoint content.

Fit and Boundaries

Flato is a good fit when the design must stay editable after generation, when AI needs to work from existing visual context, or when a presentation-centered workflow needs to expand into related visual formats.

Flato is not a one-shot image generator.

Flato is not a presentation-only tool.

Flato is not a backend development environment.

Flato is a Native Canvas Runtime for AI Design where AI agents and humans continuously co-create editable visual content.