AgentBrush icon

AgentBrush

AgentBrush is an image-generation platform for coding agents. It lets supported MCP clients create branded visuals, edit regions, and save outputs directly into a project workflow.

AgentBrush

AgentBrush for coding-agent image generation

AgentBrush is an image-generation platform built for coding agents. It exposes image tools through an MCP server so agents can create branded visuals directly from within supported CLIs and editor-based workflows.

The product is positioned for teams and builders who want project images to match an existing brand without manual prompting or constant design decisions. The site emphasizes persistent visual taste, on-brand outputs, and a workflow that stays inside the agent session.

Everything your agent needs

Preset-driven image styles

Choose from realistic photos, flat illustrations, pixel art, isometric scenes, and logos, with presets tuned for consistent output.

Brand identity inputs

Upload brand colors, fonts, and reference images so generated assets stay aligned with an existing visual identity.

Two-model generation pipeline

Draft images with a fast model, then refine them with a premium model for higher-quality results.

Mask-based edits

Paint over the area you want to change so the agent can inpaint specific regions instead of regenerating the whole image.

Multi-CLI support

Use AgentBrush through Claude Code, Cursor, Windsurf, or any other MCP-compatible client.

Practical ways teams use AgentBrush

  • On-brand product visuals

    Generate illustrations, photos, or logos while building a product so the images match the rest of the interface and brand system.

  • In-editor asset generation

    Ask the agent to create or refine assets directly in Claude Code, Cursor, or another MCP client without switching to a separate design tool.

  • Region-level image edits

    Use the mask editor to adjust only part of an image when a background, object, or local detail needs a targeted change.

  • Style-consistent asset sets

    Create output in styles such as pixel art or isometric scenes when a project needs a consistent visual system across multiple assets.

  • Agent-driven content iterations

    Let an agent handle image generation, background removal, and follow-up edits in the same workflow after the initial setup.

Pros and Cons

Pros

  • Works inside MCP-compatible agent clients instead of requiring a separate image app.
  • Supports multiple visual styles, including photos, illustrations, pixel art, isometric scenes, and logos.
  • Uses brand inputs such as colors, fonts, and reference images to help keep outputs consistent.
  • Includes mask editing for targeted refinements instead of full regenerations.
  • Documents a local login flow and project-folder output path, which fits editor-based workflows.

Cons

  • The pricing page itself is currently a placeholder rather than a live plan comparison page.
  • The public docs are described as a preview, so some setup and API details are still incomplete on the site.

FAQ

What is AgentBrush?

AgentBrush is a visual-asset generation platform for AI coding agents. It exposes an MCP server so an agent can request branded images while building an app.

Which clients does it support?

The site says AgentBrush works with any MCP-compatible client, and it is tested against Claude Code, Cursor, and Windsurf. The install guide also mentions Codex CLI and Gemini CLI.

How much does it cost?

The FAQ says plans start at $9/month for Starter with 100 tokens and scale to Power at $29/month with 1,000 tokens. The install guide shows a more detailed pricing shape on the site, but the pricing page itself is currently a placeholder.

Where do generated images go?

Generated assets are delivered to your R2 bucket and served through signed CDN URLs, with temporary artifacts cleaned up after the session ends.

Is there a free trial?

You can start with a free account and a small token allocation, then subscribe when you are ready to generate more images.

Quick Facts

Category
Developer Tool
Primary use
Image generation for coding agents
Supported clients
Claude Code, Cursor, Windsurf, Codex CLI, Gemini CLI, and other MCP-compatible clients
Source domain
agentbrush.dev
Pricing shape
Paid plans with a free account start; the FAQ cites Starter from $9/month and Power up to $29/month, while the pricing page is currently unavailable
Delivery
Generated assets are stored in an R2 bucket and delivered via signed CDN URLs

Alternativas a AgentBrush

PixelDrive icon

PixelDrive

PixelDrive is a bulk image generation API for creating many on-brand images from one template through a browser editor, REST API, CSV batch, or MCP. It suits teams that need repeatable image output with cacheable renders and usage-based pricing.

ByteAsk icon

ByteAsk

ByteAsk is a terminal-first AI coding agent for C and C++ that edits repositories and verifies changes with the real compiler, debugger, sanitizers, and tests before showing a diff. It offers a free tier plus paid plans, with editor connectors and zero-retention handling described in the source.

PXZ AI icon

PXZ AI

Una plataforma de IA todo en uno que combina herramientas para imagen, video, voz, escritura y chat para mejorar la creatividad y la colaboración.

hob icon

hob

hob is an independent workspace for coding agents that keeps agent sessions, terminals, history, and follow-up work organized around the tools and providers you already use. It is aimed at developers who want local control over routing, history, and workspace structure rather than a bundled model stack.

Manta AI icon

Manta AI

Manta AI is an autonomous web app testing tool for teams that want to map application behavior, catch regressions, and generate tests without writing scripts or maintaining selectors. It works from a URL and supports plain-English test flows, run results with screenshots, and scheduled or deployment-triggered checks.

Redline icon

Redline

Redline is a budgeting tool for Claude Code that paces sessions to stay within time, token, cost, or plan-percentage limits. It uses Claude Code’s native hooks and statusline to help sessions finish with a usable result instead of stopping abruptly.