DesignerBox MCP connects your DesignerBox account to Claude, Cursor, ChatGPT, and any other Model Context Protocol client. Once connected, you ask your assistant for product shots, video ads, on-model images, or avatars, and it generates them with your brand profile applied, then saves them to your library. 43 tools, four resources, and four named workflows, all from the chat window.
You already describe the work in a chat window. You write the brief there, argue about the angle there, and then you leave, open another tab, log into a design tool, rebuild the same context from scratch, and paste the result back.
That round trip is the thing MCP removes. Your assistant stops being where you plan the campaign and starts being where you make it.
Key Takeaways
- DesignerBox MCP is live at the server address mcp.designerbox.ai/api, with 43 tools, four auto-loaded resources, and four named workflows.
- It works with Claude Desktop, Claude.ai, Claude Code, Cursor, ChatGPT, and MCP Inspector. Adding a custom connector in Claude works on the Free plan, capped at one connector (support.claude.com, July 2026).
- Your brand profile travels with the request. The assistant reads your active brand profile before it generates, so output comes back on brand rather than generic.
- Read tools work on every plan. Generation tools need a paid plan and spend AI credits. An image is 5 credits, an avatar set of 9 images is 25.
- Video is the expensive one. A Veo 3.1 clip at 8 seconds with audio costs 6,400 credits, which is more than the Premium plan’s entire monthly allocation of 2,500. Budget it separately.
- Two destructive tools always confirm first, and the four resources are read-only, so a connected assistant cannot quietly delete your work.
What MCP is, in one paragraph
The Model Context Protocol is an open standard for connecting AI assistants to the systems where your work lives. Anthropic open-sourced it in November 2024, and in December 2025 it was donated to the Linux Foundation’s Agentic AI Foundation, which Anthropic co-founded with Block and OpenAI (anthropic.com, July 2026). More than 10,000 public MCP servers were active at that point, with support across ChatGPT, Cursor, Gemini, Microsoft Copilot, and VS Code (anthropic.com, July 2026).
The practical version: MCP is the reason one connector works in five different assistants instead of needing five different integrations.
What you can make from a chat window
Every DesignerBox model and app is reachable through the connector. You describe the result. Your assistant picks the engine.
| You ask for | What comes back | Model example |
|---|---|---|
| Studio product shots | Finished on-brand images in about 30 seconds | Nano Banana Pro |
| A launch video from a product URL | A ready-to-post clip with motion and audio | Veo 3.1 |
| A garment on a model | Editorial-ready on-model shots | Nano Banana Pro |
| A reusable brand avatar | An avatar saved by name, reusable across shots | Nano Banana Pro |
| A still animated with a camera move | A short cinematic clip | Seedance 2.0 |
| A brand audit of recent work | A read-only report flagging drift | No generation |
The catalog behind it is 13 models across six providers, and the one-click apps are available to your assistant the same way. Flat lays, product ads, try-on, retouch, background removal, photo angles: each one is a message rather than a tab.
Who this is for
Marketing teams who plan campaigns in a chat thread and want the assets produced in the same thread, without briefing a designer for every variant.
Ecommerce and DTC brands filling a catalog. One accurate product photo goes in, and the angles, backgrounds, and lifestyle scenes come out, without booking a studio day per drop. The cost comparison is worth running against real studio rates, which we broke down in what a product photoshoot costs.
Agencies running many clients, where brand profiles keep each client’s look separate and named workflows rerun the campaign that worked.
Solo founders with no design team, who want the fastest path from a product photo to something shippable.
Developers and technical teams driving DesignerBox from Claude Code, a script, or CI, rather than from a canvas.
The 43 tools, grouped the way you work
The tools are grouped by job, not by API surface. Read tools work on every plan. The generation groups spend credits and need a paid plan.
| Group | Tools | What it covers |
|---|---|---|
| Designs & search | 4 | List, fetch, and full-text search your designs and folders |
| Image & video generation | 7 | Create images, submit async video jobs |
| Avatars | 5 | Generate and manage reusable AI avatars |
| Brand profiles | 5 | Read and patch brand voice, themes, niche, assets |
| Assets, packs & templates | 7 | Browse uploads, photo packs, templates, stock |
| Pipelines & workflows | 5 | Read pipelines, community workflows, skills |
| Whiteboards & chats | 5 | Read whiteboards, galleries, assistant history |
| Teams, plan & discovery | 5 | Switch teams, read plan, credits, tool catalog |
| Total | 43 |
Five tools spend AI credits. Two tools are destructive, cancel avatar and cancel video, and both confirm before they run.
What your assistant knows before you ask
MCP has three primitives, and DesignerBox uses all three. Tools are functions the model decides to call. Resources are read-only data the application pulls in for context. Prompts are templates you invoke deliberately, which clients typically surface as slash commands (modelcontextprotocol.io, July 2026).
Four resources load at the start of a chat, with no permission prompt, because they only read:
- designerbox://account/summary for plan, credits, design count, active team, folders
- designerbox://brand-profiles/active for your brand voice, tone, themes, niche
- designerbox://teams/current for the team this connection is scoped to
- designerbox://recent-designs for your 10 most recent designs
This is the part that makes output land on brand. The assistant is not guessing your palette. It read your active brand profile before it generated anything.
Four named workflows sit in the slash menu and orchestrate several tools at once:
- /weekly-design-batch proposes weekly themes from your brand and recent work
- /audit-gallery flags recent designs that drift from your brand, read-only
- /brand-sync surfaces stale brand fields and suggests updates you approve
- /regenerate-with-style recreates an existing design in a different style
How to get started
The connector takes about a minute in a Claude client.
1. Copy the server URL.
https://mcp.designerbox.ai/api2. Add the connector. In Claude Desktop or Claude.ai, open Customize, then Connectors, click Add custom connector, name it DesignerBox, and paste the URL. Custom connectors work on Free, Pro, Max, Team, and Enterprise plans, though Free is limited to one (support.claude.com, July 2026).
3. Sign in once. Approve the DesignerBox sign-in. Your client receives a scoped 30-day token. Then you ask it to design.
For other clients:
| Client | How you add it | Notes |
|---|---|---|
| Claude Desktop / Claude.ai | Customize, then Connectors, then Add custom connector | Free plan works, one connector max |
| Claude Code | The mcp add command with an http transport and the server URL | See the docs |
| Cursor | Add the server URL to your mcp.json, global or per project | cursor.com/docs/mcp, July 2026 |
| ChatGPT | Enable Developer mode in Settings, then add the server | Pro, Plus, Business, Enterprise, Education, on the web (developers.openai.com, July 2026) |
| MCP Inspector | Point the inspector CLI at the server URL | For testing the connection |
The full connect flow for every client is on the DesignerBox MCP page.
What it costs
Credits are the unit, and they behave the same through MCP as they do in the app.
| Operation | Credits |
|---|---|
| Generate or edit an image | 5 |
| Create an avatar (9 images) | 25 |
| Brand storyboard | 10 to 25 |
| Video | Credits per second, times duration |
Plans, as of July 2026: Free is 112 credits with no credit card. Basic is $15 a month for 500 credits. Pro is $35 for 1,000. Premium is $75 for 2,500. Ultra is $200 for 8,000. Full detail is on the pricing page.
Say the video math out loud, because it surprises people. A Seedance Pro Fast clip at 720p for 5 seconds is 150 credits. Kling Standard at 720p for 5 seconds is 225. Sora 2 at 720p for 8 seconds is 1,600. Veo 3.1 with audio at 8 seconds is 6,400, which is more than a Premium plan holds in a month. Images are cheap and video is not, and no connector changes that.
What it does not do yet
Worth stating plainly, so nothing surprises you after you connect.
- Generation needs a paid plan. Read tools work everywhere, but the free tier does not generate through MCP.
- AI video and try-on clothes start at Premium ($75 a month). They are not on Basic or Pro.
- Team collaboration, shared brand kits, white label, and API access are Ultra only ($200 a month). Plans below Ultra are single seat.
- ChatGPT needs a paid plan and Developer mode, and OpenAI’s own docs describe that mode as powerful but risky, with write actions requiring confirmation (developers.openai.com, July 2026).
- Video is async. Your assistant submits the job and checks status. It does not block the chat while a clip renders.
For the wider picture on handing production to an assistant, see what to hand an AI creative agent and what to keep.
Is it secure?
The connection uses OAuth. You approve the DesignerBox sign-in once and your client holds a scoped 30-day token, rather than you pasting a long-lived key into a config file. The MCP specification requires OAuth 2.1 with PKCE for remote servers, requires tokens on the authorization header rather than in the URL, and requires servers to validate that a token was issued for them (modelcontextprotocol.io, July 2026).
Beyond the protocol: the four resources are read-only, the two destructive tools confirm before running, and your generations stay private to your account. DesignerBox does not train models on your inputs.
One honest note about the category rather than the product. A connected assistant can act on text it reads, so treat anything a tool pulls in from outside as untrusted, and read the tool input before approving a write. OpenAI says the same thing about Developer mode in its own documentation (developers.openai.com, July 2026).
FAQ
What is DesignerBox MCP?
DesignerBox MCP is a remote Model Context Protocol server that exposes your DesignerBox account to AI assistants. It publishes 43 tools, four read-only resources, and four named workflows, so an assistant like Claude can generate images, video, and avatars, read your brand profile, and search your design library from the chat window.
Which AI clients work with it?
Claude Desktop, Claude.ai, Claude Code, Cursor, ChatGPT, and MCP Inspector. Any client that supports remote MCP servers over HTTP can connect. In Claude you add it under Customize, then Connectors, then Add custom connector, which is available on Free, Pro, Max, Team, and Enterprise plans, with Free limited to one custom connector (support.claude.com, July 2026).
Do I need an API key?
Not for the Claude connector flow. You approve a sign-in once and your client receives a scoped 30-day token. API keys are managed separately in your account for scripted and CI use.
Does it use my brand profile?
Yes. Your active brand profile loads as a read-only resource at the start of a chat, so the assistant reads your voice, tone, themes, and niche before it generates. That is why output comes back on brand instead of generic.
How do credits work through MCP?
Exactly as they do in the app. Generating or editing an image is 5 credits, an avatar set of 9 images is 25, and video is priced per second of output. Read tools do not spend credits. Generation tools require a paid plan.
How long does an image take?
About 30 seconds for images. Video is submitted as an async job, so your assistant hands off the request and reports status rather than blocking the conversation.
Can it delete my designs?
No. The four auto-loaded resources are read-only. Only two tools are destructive, cancel avatar and cancel video, and both confirm with you before they run.
Which models can it reach?
The full DesignerBox catalog of 13 models across six providers, including Nano Banana Pro, GPT Image 2, Seedream 5, and FLUX 2 Flex for images, and Veo 3.1, Sora 2 Pro, Seedance 2.0, and Kling 2.6 Pro for video. Your assistant picks the engine unless you name one.
MCP protocol facts, client setup paths, and plan requirements verified against modelcontextprotocol.io, anthropic.com, support.claude.com, code.claude.com, cursor.com and developers.openai.com as of July 2026. DesignerBox tool counts, credit costs, and plan allocations verified against live product configuration and the DesignerBox MCP documentation, July 2026. Model specifications change frequently, so verify current limits against provider documentation before relying on them. Individual results vary.
