OPML Podcast & RSS Parser MCP for AI. Turns messy XML feeds into clean, structured data.

Listen up. If your AI client needs to recommend content—whether it’s the next killer podcast or an article you gotta read—it can't just stare at a dump of XML tags. You know the drill: you export your entire library from Apple Podcasts, Pocket Casts, Feedly, or whatever reader you use.

What you get back is messy. It's recursive garbage; it's verbose and bloated with tags that mean nothing to an agent trying to parse data.

That’s where this server comes in. This isn't some fancy piece of fluff designed to impress your grandma. It's a straight-up parser built for agents. Its job is simple: take that gnarly .opml file and flatten the entire thing into something usable—a clean JSON array.

The key here is data structure. LLMs don’t read nested XML trees like a human does; they choke on them. They get lost in attributes and namespaces. This server cuts through all that noise, stripping out everything extraneous so your agent can actually focus on the subscription titles and feed URLs it needs to know.

When you use parse_opml_feeds, you give it the path to your exported OPML file. It doesn't waste time trying to guess what you want; it just spits out a straightforward JSON list of every single podcast or RSS feed subscription it finds in that export. That’s it. A simple, flat array.

Think about what happens when you use this data. Instead of having your agent struggle with deeply nested XML structures—the kind where one tag needs three levels of indentation just to define a date—you give it clean, predictable JSON. The output is a list that reads like a shopping receipt: 'You subscribed to X,' 'You subscribed to Y,' and so on.

It's crucial for agents because they need reliable input. If the data format changes or gets too complex, your workflow breaks down. This server guarantees it won't. It takes the raw output from any major podcast client—the ones that generate those beastly OPML files—and standardizes the payload immediately.

The process is straightforward: you provide the file path to parse_opml_feeds, and the tool returns a clean JSON list containing every single subscription found in the source feed. This means your agent can instantly get a complete inventory of your content interests without ever having to parse or understand the underlying, messy XML markup.

It’s local too, which is huge for privacy-conscious folks like us. Your whole subscription history stays parsed locally; nothing gets sent out to some public server farm. You run this on your end, you get clean data back to your agent, and it's done. It means that when you need to build a specialized tool—maybe one that curates content recommendations based solely on the feeds you follow—you don't have to write a complex XML parsing routine yourself.

The server handles the dirty work of flattening the hierarchy.

This capability is essential for turning your personal listening habits into actionable data. You get back a simple JSON array listing all subscriptions, perfect for feeding directly into an AI agent that needs to process this list—maybe it's tracking metadata, checking feed health, or just building out a consolidated content map.

The tool doesn't care if the original file came from Pocket Casts or iTunes; as long as it follows OPML standards, parse_opml_feeds gets you clean data.

It’s about transforming that verbose XML bloat into an agent-ready list of feeds. Your AI client needs a simple array of strings and objects, not nested tags. This server gives it exactly that. It's the necessary pre-processing step for any sophisticated content curation workflow your agent might run.