# Magic Patterns MCP connector

Connect Magic Patterns to Elaichi and Claude, ChatGPT, Cursor, or the Elaichi Agent can create designs and slide decks, check design status, manage design systems, and read or write artifact files, all inside each person's own access.

Source: https://elaichi.ai/connectors/magic-patterns/

## Facts

| | |
| --- | --- |
| Application | Magic Patterns |
| Category | Design |
| AI tools | 23 |
| Authentication | Connects over OAuth |
| Bring your own OAuth app | No |
| Native MCP | Yes. Magic Patterns builds and runs this MCP server. Elaichi adds sign-in, access controls and an audit log on top |
| Support for its tools | support@magicpatterns.com |
| MCP endpoint | https://api.elaichi.ai/mcp |
| Works with | Claude, ChatGPT, Cursor, any MCP client, and the Elaichi Agent |
| Tools advertised by name | No. Connected tools are never listed one by one, however few there are. The endpoint advertises `search_tools` and `execute_tool` instead |

## What you can ask once Magic Patterns is connected

- Create a pricing page design using our main design system.
- What is the status of the onboarding flow design?
- Make a five slide deck for the Q3 product launch.

## Connect Magic Patterns in Elaichi

This happens once for the organization, before any client is involved.

1. Open Connections, choose Add connection, and pick Magic Patterns.
2. Optionally set Share with, then press Connect.
3. Approve it in Magic Patterns. Magic Patterns's own window opens. Whoever approves it decides what this connection can reach.

Credentials are vaulted and nobody, including the AI, reads them back. The connection becomes a toolbox immediately, so you can curate which Magic Patterns tools are exposed, rename them, or freeze arguments before anyone points a client at it.

## Magic Patterns MCP connector for Claude

Endpoint: https://api.elaichi.ai/mcp

1. Open Customize, then Connectors.
2. Press Add.
3. Name it, paste the MCP server URL, then Continue.
4. Sign in and approve.

On Team and Enterprise, an Owner adds it once. Everyone else turns it on for themselves.

## Magic Patterns MCP connector for ChatGPT

Endpoint: https://api.elaichi.ai/mcp

1. Open Plugins, then press the + button.
2. Name it and paste the endpoint into Server URL.
3. Leave Authentication on OAuth, then tick the risk acknowledgement.
4. Press Create, then sign in and approve.

Works on the web today. The plugin directory lives at chatgpt.com/plugins.

## Magic Patterns MCP connector for Cursor

Endpoint: https://api.elaichi.ai/mcp

1. Open `~/.cursor/mcp.json`.
2. Add the endpoint under `mcpServers`.
3. Reload Cursor, then sign in and approve.

Set up per machine, so repeat it on each computer you work from.

## Connect Magic Patterns to any MCP client

Endpoint: https://api.elaichi.ai/mcp

1. Add the endpoint as a remote MCP server.
2. Sign in and approve.

The Elaichi Agent already has these tools, with nothing to set up.

## What the consent screen decides

Only Read is granted by default, which is not enough to call a Magic Patterns tool. Over MCP there is no trusted place to confirm a write in the moment, so the consent screen is the standing approval rather than a formality. Grant Read and Run tools. Think hard before granting Delete, which reaches into connected apps and cannot be undone.

## What teams do with Magic Patterns through Elaichi

### Turn a feature brief into a first design

Product. Paste the brief and ask for a screen built on the team's design system. Check the status while it generates and send a follow-up prompt to adjust it.

### Keep every new screen on the design system

Design. Ask which design systems exist, read their files, and have new designs created against the right one. Publish updates to the design system once the team signs off.

### Draft a landing page from a campaign brief

Marketing. Describe the campaign and get a landing page design back, then iterate with short prompts until the copy and layout match the launch.

### Build a pitch deck without leaving the chat

Sales. Ask for a slide deck for a specific prospect and get a draft ready to refine, with the version history there if you need to go back.

### Pull generated files straight into the codebase

Engineering. Read the files behind an artifact, write changes back, and publish when it is ready to share with the rest of the team.

### Explore several directions before you commit

Founders. Create an inspiration document, add a few variants of the same idea, and update or clear them as the direction becomes clear.

## Elaichi vs Zapier MCP vs Composio for Magic Patterns

All three can connect Magic Patterns to an AI assistant, and all three have admin controls. They differ in where access lives and how you pay.

| What to check | Elaichi | Zapier MCP | Composio |
| --- | --- | --- | --- |
| Where the AI connects | One address for the whole organization. Endpoint: https://api.elaichi.ai/mcp | A server per member, created at sign-in. | An MCP endpoint per team, or an SDK. |
| Control over Magic Patterns tools | Allow or restrict single Magic Patterns tools, per role or user. | App and action restrictions on the account. | Role permissions, down to the action. |
| Record of calls | One audit entry per Magic Patterns call. | A History tab of tool calls. | A log of every tool call. |
| Single sign-on | SAML or OIDC, plus SCIM, on Gold. | SAML on Enterprise. | SAML and OIDC on Enterprise. |
| Price | $15 per user per month. | 2 tasks per successful call. | Billed per tool call. |

Sources: Zapier MCP [docs](https://docs.zapier.com/mcp/get-started/quickstart), [security](https://docs.zapier.com/mcp/manage/security), [usage](https://docs.zapier.com/mcp/features/usage); Composio [docs](https://docs.composio.dev/docs/composio-connect), [gateway](https://composio.dev/mcp-gateway), [enterprise](https://composio.dev/enterprise), [pricing](https://composio.dev/pricing). Checked September 2026.

Longer take: [Zapier MCP alternative](/blog/zapier-mcp-alternative/) and [when you don't need an MCP gateway](/blog/when-you-dont-need-an-mcp-gateway/).

## Frequently asked questions

### How do I connect Magic Patterns to Claude?

Connect Magic Patterns in Elaichi first, where you sign in to Magic Patterns over OAuth and approve access, with no OAuth application to register and no client ID or secret to generate. Then in Claude open Customize, then Connectors, then Add, and paste https://api.elaichi.ai/mcp. Claude signs you in to Elaichi and Magic Patterns is ready to use.

### Does Magic Patterns work with ChatGPT and Cursor as well as Claude?

Yes. Once Magic Patterns is connected in Elaichi, the same endpoint, https://api.elaichi.ai/mcp, works in Claude, ChatGPT, Cursor, any other MCP client, and the Elaichi Agent. You connect Magic Patterns once and every client picks it up.

### What can an AI agent actually do with my Magic Patterns data?

An agent connected to Magic Patterns can create designs and slide decks from a description, check whether a design is finished, send follow-up prompts, and read the recent message and version history. It can also list and read your design systems, read or write the files behind an artifact, publish artifacts and design systems, and manage variants in an inspiration document. Short, concrete asks such as a single screen with a named design system work better than long paragraphs.

### Does connecting Magic Patterns give the AI access to every design in my workspace?

No. Every call to Magic Patterns runs as the person who signed in, so the agent sees only the designs, design systems, and artifacts that person can already reach in Magic Patterns. Elaichi can narrow that access further with roles and restrictions, and it never widens it.

### Can my team share one Magic Patterns connection?

Yes. One person connects Magic Patterns in Elaichi and shares the connection with a team, and nobody else ever handles a credential. Each teammate still signs in to Elaichi as themselves, so the audit log names the person who created or published each design.

### Can I stop an agent from publishing or changing things in Magic Patterns?

Yes. Restrictions in Elaichi apply per action, so you can allow reading designs and design systems while blocking writes to artifact files or publishing. A restricted Magic Patterns action is never advertised to Claude, ChatGPT, or Cursor, so no prompt can reach it.

### What happens to a Magic Patterns connection when someone leaves?

Offboarding a person in Elaichi ends their access to Magic Patterns at once, across every client they used. A connection shared with a team keeps working for everyone else. Disconnecting Magic Patterns once in Elaichi removes it from Claude, ChatGPT, Cursor, and every other client at the same time.

### Does the Magic Patterns MCP connector work with Gemini, Codex, Claude Code or other MCP clients?

Yes. Magic Patterns is reached over the same MCP endpoint every client uses, so anything that speaks MCP can call it — Gemini, Codex, Claude Code, Windsurf, Cline, Zed and OpenCode among them — alongside Claude, ChatGPT, Cursor, and the Elaichi Agent. The tools on offer and the access behind them are identical whichever client asks. Only the setup screen differs.

### Is Elaichi an alternative to Zapier MCP for Magic Patterns?

Yes. Both let Claude, ChatGPT or Cursor use Magic Patterns. Zapier MCP fits a team that already automates in Zapier, since each person signs in and acts as themselves in that account. Elaichi fits when IT wants one address for the whole company, per-tool rules by role, and a record of every Magic Patterns call.

### How is Elaichi different from Composio for Magic Patterns?

Composio gives AI agents tools and sign-in handling across 1,000+ apps, for developers building agents or people using an assistant, billed per tool call. Elaichi gives a company's own people governed access to Magic Patterns: one address, restrictions per role or user, and $15 per user per month. Both have role permissions and a log of every call.

## All 23 Magic Patterns tools

Every tool below is callable through https://api.elaichi.ai/mcp once Magic Patterns is connected, subject to the toolbox it is in and the restrictions on the caller.

- **List design systems** (List). Lists the design systems available to the authenticated user. Returns both built-in presets (e.g. Base, Shadcn, MUI) and any custom design systems.
- **Create design** (Create). Creates a new Magic Patterns design. This is the starting point for both workflows.
- **Create slide deck** (Create). Creates a new Magic Patterns slide deck and kicks off AI generation. A slide deck is a 16:9, full-bleed, one-slide-at-a-time React presentation where each slide maps to a screen in the canvas. Unlike `create_design`, a `prompt` is required — there is no blank-deck path.…
- **Get editor ID from URL** (Get). Resolves a Magic Patterns URL to an editor ID.
- **Get design status** (Get). Gets the current status of a design: whether AI generation is active, the active artifact ID, and available files.
- **Send prompt** (Send). Sends a natural language prompt to the Magic Patterns AI for an existing design. Returns immediately with a request ID. The generation runs in the background.
- **Read recent message history** (Action). Reads the recent chat item history for a design. Returns the last 10 items (user prompts, AI responses, artifact versions, edits). Use `skip` to paginate backwards. Code contents are omitted to keep the response concise — use `read_artifact_files` to read full file contents.
- **List version history** (List). Lists the artifact version history for a design. Returns the most recent 20 versions.
- **Publish artifact** (Publish). Compiles an artifact's source files and sets it as the active artifact for the design. This is the final step in the code-first workflow.
- **Get artifact** (Get). Gets the active artifact for a design, including its ID and list of files. Always call this (or `get_design_status`) to get the latest active artifact — do not rely on a previously cached artifact ID.
- **Create new artifact** (Create). Creates a new artifact by cloning an existing one. The new artifact becomes the active artifact.
- **Read artifact files** (Action). Reads the contents of one or more files from an artifact.
- **Write artifact files** (Action). Creates or overwrites one or more files in an artifact. If a file exists, it is replaced. If it doesn't exist, it is created.
- **Get design system** (Get). Resolves a design system's active artifact and lists its files. Always call this first — design systems are collaborative, so the active artifact ID can change between calls. Never reuse a cached `artifactId`.
- **Read design system files** (Action). Reads the contents of one or more files from a design system's active artifact.
- **Create design system** (Create). Creates a new, blank design system owned by you and returns its ID and editor URL. Seeds an empty initial version so you can immediately write files into it.
- **Write design system files** (Action). Creates or overwrites files in a design system. Incoming files are merged onto the existing artifact (existing files are preserved), then compiled and activated immediately.
- **Publish design system** (Publish). Publishes the design system's active artifact as a new immutable version.
- **Create inspiration document** (Create). Creates an inspiration document and returns a shareable URL. Declare each concept's `name`/`description` without `html` to publish the link immediately, then stream each concept in with `inspiration_add_variant`. You may also pass `html` inline to publish concepts fully in one…
- **Inspiration add variant** (Add). Fills in one placeholder concept with its content. The concept renders live on the shared page as soon as its content arrives; the document flips to "ready" once every concept is filled.
- **Get inspiration document** (Get). Loads an inspiration document by its ID. Use this to read concepts back.
- **Inspiration update variant** (Update). Revises a single already-filled concept in place, replacing its content (and optionally its name/description). Use this to update a subset of concepts without touching the others.
- **Inspiration clear variants** (Clear). Resets every concept back to an empty placeholder — dropping each concept's content and its pre-created "Iterate" room. Use this to replace all concepts with a fresh set of directions.