Pletor MCP: orchestrate your workflows with agents (Claude, Codex, ...)

Creative demand is scaling. Creative infrastructure does not.

The workflows that made your last campaign work aren't easily operable with the rest of your stack. The brief is in your colleague’s worskpace. The performance data sits in a spreadsheet your agent can't read. Every launch means connecting all these sources manually.

Today we're closing one of those gaps. The Pletor MCP makes every workflow you've built operable from any MCP-compatible agent — Claude, Codex, or whatever your team already runs on — and connectable to the tools around it.

Workflows stop being islands. They run alongside your stack.

‍

How it works

Claude, Codex, or any MCP-compatible agent becomes the entry point. You describe what you want: the inputs, the brief, the scale, the workflows.

The agent pulls the context it needs, shapes the parameters, and triggers the right Pletor workflows.

Then it runs them. Once, or a hundred times.

‍

What it unlocks

The same architecture supports any creative production loop an agent can describe. Three concrete examples:

Performance-informed creative. Claude reads your winning Meta ads from the last quarter, extracts the patterns that worked, and feeds that context into your Pletor Meta Ads workflow. The next campaign starts where the last one left off.

Pull, run, push. Claude pulls dozens of products from Google Sheets or Airtable, runs your PDP Pletor workflow, and pushes the finished assets back to Shopify. The marketing lead never opens three tools to ship one drop.

From product page to video. Claude scrapes a product page, extracts benefits and pack shots, and runs a Pletor workflow that turns each SKU into a promotional video. One scrape, one prompt, one folder of assets.

‍

Why it compounds

Most MCPs in the creative space let an agent call a model. That's useful, but it's a single-axis operation. The agent generates faster; the production system still lives in someone's head, a Notion doc, or a tab nobody updates.

Pletor's MCP runs the production system itself. The workflow carries the context. The agent operates the orchestration. Every workflow plugs into the tools around it.

A team that encodes its production into Pletor workflows builds a system that gets faster, more consistent, and more callable over time, from any surface, by any agent, alongside any tool. Brand knowledge becomes infrastructure.

‍

Run it

Add Pletor as a custom MCP server in Claude:

{
  "mcpServers": {
    "pletor": {
      "url": "https://api.pletor.ai/mcp"
    }
  }
}

‍

Authenticate once. From that point forward, every Pletor workflow on your account is callable from your agent.

Full setup guide: docs.pletor.ai/automate/pletor-mcp →

‍

Creative production is becoming agent-driven. We're building the infrastructure underneath it.

‍