Cadru with your AI assistant
Tell it what you want - from a link, photos or an idea - and the carousel lands in your editor. It can also change the open one while you watch: “make the title bigger”, “put this photo on slide 3”.
Tell it what you want - from a link, photos or an idea - and the carousel lands in your editor. It can also change the open one while you watch: “make the title bigger”, “put this photo on slide 3”.
In Claude: Settings, Connectors, Add custom connector - with the address below.
cadru.photo opens: your email, the code you get, Allow. A new email gets an account on the spot.
From photos, an article or an idea. It lands in your editor and you can change it any time.
https://cadru.photo/mcp/carouselConnect to Cadru so you can make carousels for me: the instructions are at https://cadru.photo/aiclaude mcp add --scope user --transport http cadru https://cadru.photo/mcp/carouselThen type /mcp, pick “cadru” and Authenticate - cadru.photo opens for your email, the code and Allow.
Add a remote MCP server with the address above and OAuth authentication. Client registration is automatic.
Photos show what happened, only you know why it matters. A few sentences are enough: what it is about, who it is for, names, places, a line of your own. It writes the words on the slides from them. If the story is missing it asks; if you are in a hurry, just say “make it”.
On your computer: with Claude Code or the desktop app, just name the folder - it sends them straight into the carousel (photos, clips, SVGs) through a private 15-minute link. On your phone: open cadru.photo/post (signed in), drop them into the Library and tell the assistant they are there - it lays the carousel out from them. Don't send them into the chat one by one.
If you are an assistant and the person asks you to connect to Cadru: the MCP server for carousels is https://cadru.photo/mcp/carousel (Streamable HTTP, OAuth 2.1 with the Cadru account, dynamic client registration). In claude.ai and the Claude apps you cannot add the connector yourself - give the person the steps above. Photos, clips and SVGs on the person's computer (if you can run commands): upload_link, then curl - only the files they asked for. The full instructions, per kind of client: https://cadru.photo/llms.txt. The server card: https://cadru.photo/.well-known/mcp.json