Guides14 Oct 20265 min read
Sloosh MCP: make images and video from Claude, ChatGPT or Cursor
Connect your AI app to Sloosh over MCP, set a daily limit, then ask in plain words. Every paid run starts with a quote.
Sloosh MCP lets the AI app you already use, such as Claude, ChatGPT or Cursor, make images and video in your Sloosh workspace. Add the server address as a connector, sign in once in your browser, set a daily limit, and ask in plain words. Your agent gets a quote before anything is charged, and the result is saved in your Sloosh Assets.
MCP (Model Context Protocol) is an open standard that lets an AI assistant use tools from other apps. Sloosh's MCP server gives your assistant tools to read your workspace, find assets, run Spaces, generate images and video, and build Spacelab canvases.
Connect your AI app, step by step
1. Open the MCP page and pick your app
Open MCP in the Sloosh top bar. Pick your app from the tabs (ChatGPT, Claude, Cursor, Claude Code and more) and the page shows its steps.
2. Copy the server address and add Sloosh as a connector
The address is the same for every app:
https://mcp.sloosh.ai/mcpIn Claude (web or desktop), open Customize, then Connectors, choose Add, then Add custom connector. Name it Sloosh and paste the address. You can also add Sloosh to Claude as a plugin, downloaded from the MCP page, which adds skills for product images, ad variants and more. In ChatGPT, turn on Developer mode in settings, create an app named Sloosh, paste the address and choose OAuth. In Claude Code, run:
claude mcp add --transport http sloosh https://mcp.sloosh.ai/mcpthen run /mcp to sign in. Cursor, VS Code, Codex and other tools are covered in the MCP docs.
3. Sign in to Sloosh in your browser
The first time, your browser opens and you sign in to Sloosh. You pick the workspace the connection works in. You never paste a password or a key into the app.
4. Set a daily limit
Nothing paid runs until a daily limit is set. On the MCP page, open Connected Agents, then Agent Settings, and choose how many credits agents can use each day. You can also tick Ask before spending more than so the agent checks with you above a set amount. Until a limit is set, the agent can read and build drafts but can't spend.
5. Ask for what you want
Write the request the way you'd say it. This is the one from the video:
Use Sloosh to put my tan tote on a striped beach towel, bright summer light.
Your agent picks a model, or one of your Spaces, and gets a quote before anything is charged. Then it runs.
6. Find the result in your chat and in Assets
The image comes back to your chat. It's also saved in your Sloosh Assets, ready to edit or share.
More prompts to try
These starter prompts are from the MCP docs. Paste one once your tool is connected and a daily limit is set:
Make a studio photo of my water bottle on a clean white background.
Create three 9:16 ad variants for my summer sale on this sneaker.
Find a space in my workspace that makes product ad grids and tell me what a run costs.
Turn my latest product image into a short video clip.
Which workspace your agent works in
When you sign in, you pick the workspace the connection works in, and it stays there. To work in another workspace, connect again and pick it. If you're not sure where you are, ask your agent to call get_connection: it shows the workspace, the permissions you granted and how much of today's limit is left.
Build a Space from your chat
Your agent can do more than generate. With the spacelab:author permission it can create a draft canvas, edit it and price a test run, then show you a publish summary before it publishes the canvas as a Space in your workspace. Only a workspace admin can publish, and the agent can't replace a Space someone made public.
What keeps spending in check
- A daily limit you set. Every AI tool connected to the workspace shares it, the agent can't raise it, and it resets at midnight UTC.
- A quote before every paid run. The agent sees the price before anything is charged, and a quote is good for 5 minutes.
- Three permissions.
spacelab:readreads your workspace,spacelab:authoredits and publishes drafts without spending, andspacelab:executespends credits. - Drafts only. An agent can build and edit canvases, but a published Space doesn't change until someone publishes again. Every agent edit saves a version, labelled with your AI tool's name, that you can restore.
To stop an app, open Connected Agents on the MCP page and disconnect it. New calls stop right away; runs already under way still finish.
Related
- Space vs Spacelab: what your agent is running when it picks a Space.
- How to keep a consistent AI character: save products and people as Elements so every run starts from the same picture.
- Agent MCP access and plans: Pricing.
Questions
- Can an AI agent spend my Sloosh credits?
- Only after someone sets a daily limit in Sloosh's MCP settings and you grant the spacelab:execute permission. Every paid run starts with a quote, and Ask before spending more than sets when the agent must check with you.
- Do I paste an API key into Claude or ChatGPT?
- No. You add the server address, and your browser opens so you sign in to Sloosh. You never paste a password or a key into the app.
- When does the daily limit reset?
- At midnight UTC. Every AI tool connected to the workspace shares the same limit, the agent can't raise it, and reconnecting doesn't reset it.
- Can an agent change a Space I already published?
- No. An agent edits drafts only, and a published Space doesn't change until someone publishes again. Every agent edit saves a version you can restore.
- Which plan includes MCP?
- The Pricing page lists agent MCP access on the Pro plan, and Max includes everything in Pro.
Try it in Sloosh
Make an image in Studio with what you just read.
