PaneFlow MCP Server - Build Slideshows with AI Agents

The PaneFlow MCP server lets your AI assistant build and edit real PaneFlow projects for you. Ask it in plain language and it creates slideshows, presentations, and decks, adds slides, positions text and images, applies animations and transitions, generates images, and more. Everything it makes is a real project in your dashboard that you can open in the editor anytime to fine-tune, present, or export.

It works with any MCP-capable client - Claude Code, the Claude desktop and web apps, Cursor, VS Code, and others.

#What you can do

Just describe what you want. For example:

Build a deck from a topic, a brief, or an outline
Add, duplicate, reorder, and edit slides
Place and style text, images, videos, shapes, charts, code blocks, and 3D device mockups
Apply animations and the block transitions that make panes flow into each other
Generate images and remove backgrounds with AI
Take screenshots of any slide so the assistant can see its work and fix issues

#Requirements

An active PaneFlow plan. The MCP server is a paid feature - you connect with the account you subscribe with. See plans.
An MCP-capable AI client, such as one of the ones below.

#Connect

The PaneFlow MCP server lives at a single URL:

https://paneflow.com/mcp

Add it to your client using the steps for your tool below. The first time you use it, the client opens your browser to sign in to PaneFlow and authorize access (see Signing in).

#Claude Code

Run this in your terminal:

claude mcp add --transport http paneflow https://paneflow.com/mcp

Then start (or restart) Claude Code and run /mcp to authenticate.

#Claude desktop & web apps

Open Settings -> Connectors -> Add custom connector, then enter:

Name: PaneFlow
URL: https://paneflow.com/mcp

Custom connectors are available on paid Claude plans.

#Cursor

Add PaneFlow to .cursor/mcp.json (in your project) or to your global Cursor MCP settings:

{
  "mcpServers": {
    "paneflow": {
      "url": "https://paneflow.com/mcp"
    }
  }
}

#VS Code

Create .vscode/mcp.json in your workspace:

{
  "servers": {
    "paneflow": {
      "type": "http",
      "url": "https://paneflow.com/mcp"
    }
  }
}

#Codex (OpenAI)

Add PaneFlow from the terminal:

codex mcp add paneflow --url https://paneflow.com/mcp

Or add it to ~/.codex/config.toml:

[mcp_servers.paneflow]
url = "https://paneflow.com/mcp"

#OpenCode

Add PaneFlow to your opencode.jsonc:

{
  "mcp": {
    "paneflow": {
      "type": "remote",
      "url": "https://paneflow.com/mcp",
      "enabled": true
    }
  }
}

#Other clients

Any client that supports remote MCP servers over HTTP works. Point it at https://paneflow.com/mcp using that client's "remote" or "HTTP" server configuration.

#Signing in

PaneFlow uses standard OAuth (2.0 with PKCE), so you never paste an API key or token.

The first time your client connects, it opens a browser window where you sign in to your PaneFlow account and approve access. After that, your client refreshes access automatically - you stay connected. You need to sign in with an account that has an active subscription.

To disconnect, remove the server from your client's MCP settings.

#How to prompt

The assistant works best with a bit of direction. A few tips:

Describe the deck up front: its purpose, audience, tone, and roughly how many slides. "A 6-slide investor pitch for a fintech startup, clean and confident, dark theme."
Share your brand: mention colors, fonts, and any logos or images it should use.
Iterate slide by slide. It's easy to refine: "make the title bigger," "swap slide 3's background for something warmer," "add a closing call-to-action slide."
Ask it to check its work. The assistant can screenshot any slide to verify layout and styling, so prompts like "take a screenshot of slide 2 and fix the spacing" work well.
Reference existing projects by name to keep editing decks you already started.

#Example prompts

Create a 5-slide product launch deck for my app "Nimbus". Start with a hero title slide, then three feature slides, then a closing call-to-action. Use a dark, modern look.

Add a new slide after slide 2 with a big stat - "10,000+ teams" - and a short caption underneath.

Animate the title on the first slide so it blurs in, and make the whole deck transition more smoothly between slides.

Generate a hero image of a mountain landscape at sunset and use it as the background of slide 1.

Open my "Q3 Review" project and turn the bullet list on slide 4 into a bar chart.

#AI images

The assistant can generate images and remove backgrounds on your behalf. These operations use AI credits from your balance, the same as in the editor. The assistant tells you the cost before it spends anything - and you can give it a budget up front, for example "feel free to generate up to 5 images."

#FAQ

Do I need to know how to code?
No. You talk to the assistant in plain language; it handles everything in PaneFlow.

Where do the slideshows go?
Straight into your PaneFlow dashboard. Open any project in the editor to tweak it by hand, present it, or export to HTML, React, Vue, Svelte, video, PDF, and more.

Is it free?
The MCP server is included with a paid PaneFlow plan. AI image generation uses credits, just like in the app.

My client won't connect.
Make sure you're signed in to PaneFlow with a subscribed account, that the server URL is exactly https://paneflow.com/mcp, and that your client supports remote HTTP MCP servers. Restarting the client after adding the server usually triggers the sign-in prompt.