AgentBrush
MCP server that gives coding agents persistent visual taste for on-brand AI image generation
If your agent already writes your code, AgentBrush closes the biggest remaining gap — visuals that don't look improvised. The brand identity layer and the draft-then-refine pipeline are what separate it from a thin image API call, and the mask editor keeps iteration cheap because you repaint one region instead of regenerating the whole asset. Multi-CLI support means one subscription covers Claude Code, Cursor, and Windsurf rather than a seat per editor. The August 2026 cMCP release, which denies an agent tool call and returns a signed receipt, signals the team is building toward agent oversight, not just image output. Skip it if you need manual pixel control or run everything locally — this
Verified 10d ago · liveness 80/100 · cite: rightaichoice.com/tools/agentbrush
- Indie developers generating on-brand visuals for AI-built websites
- Game devs who need pixel art or isometric sprites inside their IDE
- Agency teams keeping marketing creatives visually consistent across projects
- Teams standardizing on Claude Code, Cursor, or Windsurf for development
- Designers who need manual, pixel-level creative control
- Teams requiring a local-only or privacy-isolated image generator
- Users who prefer a drag-and-drop GUI over CLI and agent workflows
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip AgentBrush if you need layer-based manual control over every pixel or you run image work in a local-only environment — this is an agent-first MCP server, not a design app.
Published tiers put AgentBrush at $0/mo for Starter and $20/mo for Pro, with a custom-priced Team plan for multiple seats and shared brand identity. That sits in the same band as a single mainstream AI image subscription, but one $20/mo Pro subscription covers every MCP-compatible agent you connect, so the effective per-editor cost drops as you add Claude Code, Cursor, or Windsurf rather than paying per seat per tool. The Team tier is quote-based, so agency pricing scales with seats rather than
In short
AgentBrush — MCP server that gives coding agents persistent visual taste for on-brand AI image generation. Best for Indie developers generating on-brand visuals for AI-built websites, Game devs who need pixel art or isometric sprites inside their IDE, Agency teams keeping marketing creatives visually consistent across projects. Free to start; paid plans from $20/mo.
What's new in AgentBrush
Checked 10 days agoAcross the latest 1 update: 1 feature update.
What people actually say about AgentBrush — is it worth it?
We ran a structured research pass across product reviews, community discussions, and post-purchase forum threads to surface the patterns vendors won't publish themselves. Below: the recurring strengths, the hidden costs people mention most, and the cohort that consistently regrets adopting this tool.
7 mentions across 2 sources (Hacker News, Product Hunt) · researched Jul 2, 2026.
Average across the 2 sources that answered — each source counts once, not each post.
- +Keeps visual asset creation inside the coding workflow.
- +Persistent brand identity storage ensures consistent on-brand outputs.
- +Generates images directly into the repo without manual downloads.
- +Two-model pipeline (draft + refine) balances speed and quality.
- +Supports multiple agents with a single subscription.
- −Only works with MCP-compatible agents, limiting audience.
- −Very limited community feedback – unknown real-world reliability.
- −No integration with non-MCP tools like Midjourney or Figma.
- −Mostly targets developers; non-coders have no use for it.
- −No transparent pricing tiers or hidden cost details.
- • No clear pricing tiers disclosed in data; usage-based billing could surprise heavy users
Viability Score
How well maintained and how widely used is AgentBrush? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: October 2026
How we score →Key Features
- MCP server that plugs image generation into coding agents
- Five generation presets: realistic, flat illustration, pixel art, isometric, logo
- Brand identity upload with colors, fonts, and reference images
- Two-model draft-refine pipeline for photorealistic output at lower cost
- Mask editor for region inpainting and iterative edits
- Multi-CLI support: Claude Code, Cursor, Windsurf, any MCP agent
- One subscription covering every connected agent
- Transparent background PNG generation
- On-brand image generation without manual prompt engineering
- cMCP: deny an AI agent's tool call and receive a signed receipt
- Audit trail for agent tool calls
- Commercial use rights
About AgentBrush
AgentBrush is an MCP server that puts image generation inside the coding agents you already run. Ask Claude Code, Cursor, or Windsurf for a logo, illustration, pixel art sprite, or marketing image and the agent runs the pipeline itself, so you don't tab out to Canva or Figma mid-build. The product is built around five primitives. Presets cover realistic photos, flat illustrations, pixel art, isometric scenes, and logos, each tuned so output stays consistent between runs. Brand identity uploads carry your colors, fonts, and reference images into every generation, removing the manual prompt-tuning that normally lets assets drift. A two-model pipeline drafts with a fast model and refines with a premium one, and a mask editor lets the agent repaint a region instead of regenerating the whole image. Multi-CLI support means one subscription across every MCP-compatible agent you use rather than a seat per tool. In August 2026 AgentBrush also shipped cMCP, which denies an agent's tool call and returns a signed receipt — useful when you need an audit trail of what your agent actually did. It targets indie devs shipping AI-built websites, game makers who need sprites fast, and agency teams keeping every creative on the same brand. Published tiers run Starter at $0/mo, Pro at $20/mo, and a custom-priced Team plan, though the pricing and release-notes pages we fetched did not render tier tables. Against manual editors the trade is speed and brand consistency for fine-grained control: favorable if your workflow already lives in the terminal, not if you need to nudge individual pixels.
Behind the Verdict
AgentBrush makes one bet and makes it consistently: the image generation step belongs inside the coding agent, not beside it. That shows up in the design. Five primitives — presets, brand identity, a two-model draft-refine pipeline, a mask editor, and multi-CLI support — are things an agent can call programmatically, not features that need a human clicking around a canvas. Presets matter more than they sound: realistic, flat illustration, pixel art, isometric, and logo each being tuned for consistency is what stops a set of ten generated assets from looking like ten different designers made them. The brand identity upload carries your colors, fonts, and reference images into every generation, which is the actual mechanism behind the on-brand promise — consistency depends on how good that upload is, so budget real time for it once and reuse it. The two-model pipeline is the cost play: draft with a fast model, refine with a premium one, and you get photorealistic output without paying premium rates on every token of the first pass. The mask editor is the iteration story, and it is the one that determines whether you actually use this or go back to Figma after the first bad logo — repainting a region is cheap, regenerating a whole image and hoping the rest stays put is not. Where AgentBrush is thinner is at the edges. The underlying models in the two-model pipeline are not named anywhere on the pages we could reach, so if you need to know what is generating your brand assets, this run cannot tell you. The mask editor is an agent primitive, not a Photoshop brush: if your work requires nudging individual pixels, a designer-grade tool is still the right answer. And because it is an MCP server rather than a standalone app, your mileage depends on your client — Claude Code, Cursor, and Windsurf are the named targets and multi-CLI support is claimed for any MCP-compatible agent, but there is no GUI to fall back on when the agent gets it wrong. The cMCP release is the most interesting thing here for anyone nervous about handing an agent write access to production assets. Denying a tool call and returning a signed receipt means you can reconstruct what your agent tried to do, which is the beginning of an audit story rather than a marketing gesture. If that direction continues, AgentBrush becomes less an image tool and more an agent-permissions layer that happens to generate images. Worth watching, and worth trying on the Starter tier if you are already living in a terminal.
Researching AgentBrush? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Real-world workflow fit
Concrete scenarios for the personas AgentBrush actually fits — and what changes day-one when you adopt it.
You are mid-build in Claude Code and need a hero illustration plus a transparent-background logo. You tell the agent what you want, AgentBrush picks a preset, pulls your uploaded brand colors and fonts, drafts with the fast model, refines with the premium one, and drops the PNG into the project.
Outcome: You get on-brand assets without opening Canva or Figma, and the brand identity upload keeps the next five images matching the first ones.
You are prototyping in Cursor and need a set of pixel art sprites, a tileset, and inventory icons that share a palette. You describe the set, AgentBrush runs the pixel art preset across each item, and the mask editor repaints a single sprite's boots when one comes back wrong.
Outcome: A consistent sprite set lands in the IDE without a separate asset pipeline, and per-item fixes stay cheap because you repaint a region instead of regenerating the sprite.
Your team runs multiple campaigns across different client brands. Each brand's colors, fonts, and reference images live as an uploaded identity, so any developer on Claude Code, Cursor, or Windsurf generates campaign creatives that stay inside the client's palette.
Outcome: Shared brand identity keeps every seat producing the same look, and cMCP signed receipts give you a record of which agent calls produced which assets.
Use Cases
- Generate on-brand marketing and social creatives at scale using your agent
- Create pixel art sprites, tilesets, and icons for 2D game development
- Design transparent-background PNG assets for websites and apps
- Generate realistic product photography without a studio setup
- Iterate on specific image regions using the mask editor for inpainting
- Produce app-store-ready icons and logos directly inside Claude or Cursor
- Audit AI agent actions with cMCP signed receipts
Limitations
- AgentBrush is an MCP server that works with MCP-compatible coding agents such as Claude Code, Cursor, and Windsurf, and it does not appear to offer a standalone interface — there is no GUI to fall back on when the agent gets a generation wrong.
- Brand consistency depends on the brand identity you upload, so the quality of your colors, fonts, and reference images is the ceiling on the output.
- The underlying AI models used in the two-model pipeline are not named on the pages we could reach, so you cannot tell from the public pages what is generating your assets.
- The mask editor is an agent-callable inpainting primitive rather than a layer-based design canvas, so fine pixel work still belongs in a dedicated editor.
as of 2026-10-01
Verification history
We have re-verified AgentBrush 6 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
Free to cite with attribution — this page re-verifies continuously.
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Plans compared
For each published AgentBrush tier: who it actually fits, and what it adds vs. the previous tier. Cross-reference the cost calculator above for projected annual outlay.
Starter
$0/mo
Ideal for
A solo developer who wants to try AgentBrush inside one coding agent before committing, and is generating a handful of icons or illustrations for a side project.
What this tier adds
Free entry point; covers the MCP server connection for supported CLIs and access to the generation presets.
Pro
$20/mo
Ideal for
Indie devs and game developers generating assets regularly inside Claude Code, Cursor, or Windsurf who need output to match an established brand across many images.
What this tier adds
Adds full access to all five presets, brand identity uploads, the two-model draft-refine pipeline, the mask editor, and multi-CLI support, all for $20/mo.
Team
Custom
Ideal for
Agencies and multi-developer teams producing creatives for several client brands at once and needing a shared visual standard across seats.
What this tier adds
Adds multiple seats and shared brand identity on top of everything in Pro; priced custom rather than at a published rate.
Where the pricing makes sense
The company stage and team size where AgentBrush's pricing actually pencils out — and where peers do it cheaper.
Published tiers put AgentBrush at $0/mo for Starter and $20/mo for Pro, with a custom-priced Team plan for multiple seats and shared brand identity. That sits in the same band as a single mainstream AI image subscription, but one $20/mo Pro subscription covers every MCP-compatible agent you connect, so the effective per-editor cost drops as you add Claude Code, Cursor, or Windsurf rather than paying per seat per tool. The Team tier is quote-based, so agency pricing scales with seats rather than
Setup time & first value
How long it actually takes to get something useful out of AgentBrush — broken out by persona, not the marketing-page minute.
Indie devs already in Claude Code, Cursor, or Windsurf: connect the MCP server and you can generate a first image in the same session. First real value takes longer because brand identity is the setup cost — uploading colors, fonts, and reference images well is what makes output consistent. Agency teams should budget additional time to define shared brand identities per client alongside their
Switching to or from AgentBrush
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From Canva: keep Canva for finished marketing layouts, and route in-IDE asset generation through AgentBrush so icons and illustrations land in the repo
- →From Figma: move logo and icon generation into your coding agent, and stay in Figma for component-level UI design
- →From ad hoc image APIs: replace hand-written prompt calls with AgentBrush presets and one brand identity so output stops drifting between runs
- →From manual prompt engineering: upload brand colors, fonts, and reference images once instead of re-describing your brand on every generation
- ↗To Canva: export your generated assets and rebuild brand controls in Canva's brand kit if you need a browser canvas and non-developer teammates
- ↗To Figma: move assets into Figma when the work shifts from generation to layer-level composition and design-system maintenance
- ↗To a standalone image generator: if you need a GUI rather than agent calls, and can accept re-establishing brand references somewhere else
Integrations
Resources & Guides
- Documentationagentbrush.dev
Docs · AgentBrush
Full product docs from agentbrush.dev
- Quickstartagentbrush.dev
Getting Started · AgentBrush
Get up and running fast from agentbrush.dev
- Resourceagentbrush.dev
Faq · AgentBrush
Helpful link from agentbrush.dev
- Resourceagentbrush.dev
Blog · AgentBrush
Helpful link from agentbrush.dev
- Resourceagentbrush.dev
Changelog · AgentBrush
Helpful link from agentbrush.dev
Tutorials & Learning
YouTube returned 6 videos for “AgentBrush”, and we withheld 6: 6 could not be judged, because “AgentBrush” is a single word that other videos use for other things. We are showing none, because we could not prove any of them are about AgentBrush.
Official links
Tools that pair well with AgentBrush
Common stack mates teams adopt alongside AgentBrush, with the specific reason each pairing earns its keep.
Runway Gen-4
Runway Gen-4: AI video generation and frame-level editing in one credit-based studio, now with plugin access inside Adobe, Resolve and MCP agents.
Lalaland.ai
Enterprise AI fashion models for Browzwear's 3D design ecosystem — 3D assets in, on-brand product imagery out.
Envato Elements
Envato Elements pairs unlimited creative asset downloads with AI video, image, audio and voice generation under one commercial license.
Featured Head-to-Head Comparisons
Agentbrush vs Cognition Ai
Choose AgentBrush if you're a coding agent user needing on-brand image generation (logos, pixel art, illustrations) integrated into Claude Code/Cursor/Windsurf — it's purpose-built for that. Choose Cognition AI's Devin if you're an enterprise team wanting an autonomous engineer that creates pull requests, runs in Windows VMs, and triages bugs, backed by a $10M guarantee. They solve completely different problems; one augments agents with visuals, the other is an agent itself.
Agentbrush vs Bito
AgentBrush and Bito solve completely different problems. AgentBrush is for generating images from within coding agents (presets, brand identity, inpainting). Bito provides system-wide context from code and docs to agents, enabling cross-repo impact analysis and automated scoping. Choose AgentBrush if you need AI-generated visuals; choose Bito if you need deep code understanding across repositories.
Agentbrush vs The New Black
Choose AgentBrush if you're a developer using Claude Code, Cursor, or Windsurf who needs to generate website visuals, pixel art, or isometric game assets straight from your coding agent. Choose The New Black if you're in fashion — designing apparel or accessories with tech pack exports and virtual try-on. They serve entirely different domains, so the decision hinges on whether your work is code-driven or fashion-centric.
Alternatives to AgentBrush
View all
Runway Gen-4
Runway Gen-4: AI video generation and frame-level editing in one credit-based studio, now with plugin access inside Adobe, Resolve and MCP agents.
Lalaland.ai
Enterprise AI fashion models for Browzwear's 3D design ecosystem — 3D assets in, on-brand product imagery out.
Envato Elements
Envato Elements pairs unlimited creative asset downloads with AI video, image, audio and voice generation under one commercial license.
Frequently Asked Questions
Best-of guides
Used AgentBrush? Help shape our editorial sentiment research.