You ask your AI assistant to summarise a long research paper. It produces three thousand words of insight. You skim the first paragraph, promise yourself you’ll come back to it, and never do.
Sound familiar?
Today we’re launching the noctua MCP server — a way for your AI assistant to send text directly into your noctua podcast feed. Ask Claude to summarise something, then ask it to send the summary to noctua. By the time you pick up your phone, it’s an edition waiting in your podcast app.
What is MCP?
MCP — the Model Context Protocol — is an open standard that lets AI assistants talk to external tools. Think of it as a universal adapter: once an AI app supports MCP, it can connect to any MCP server and use its capabilities. Claude Desktop, Cursor, and a growing list of AI tools all speak it.
noctua is now one of those tools. Connect it once, and your AI assistant gains three new abilities:
convert_text_to_audio— send any text or markdown and it becomes an edition in your feed.list_episodes— see what’s already been queued up.get_episode_status— check whether a conversion is ready.
What this unlocks
The interesting part isn’t the protocol — it’s what becomes possible when your AI assistant can hand work off to noctua directly.
- “Summarise this 80-page PDF and send it to my podcast.” Your assistant condenses, noctua reads it aloud. Listen on your walk home.
- “Take this long Slack thread and turn it into a briefing edition.” Catch up on what happened while you were in meetings — without scrolling.
- “Here’s my reading list. Turn each article into an edition.” One prompt, a full queue.
- “Research [topic] and send me a fifteen-minute summary as a podcast.” Your assistant does the gathering and the writing; noctua does the narration.
The pattern: anywhere your AI produces or processes text you’d rather hear than read, send it to noctua and move on.
Setting up Claude Desktop
This is the bit that takes a minute. After this, you never think about it again.
Step 1: Open Connector settings
In Claude Desktop, open Settings → Connectors. You’ll see a list of any connectors you already have configured.
Step 2: Add noctua as a custom connector
Scroll to the bottom of the Connectors list and click Add custom connector. In the dialog that appears, paste the noctua server URL:
https://api.noctua.uno/mcp
Leave the advanced settings alone — noctua handles OAuth for you — and click Add.
Step 3: Sign in to noctua
Claude will open a browser tab and ask you to sign in to noctua — the same account you use at app.noctua.uno. Sign in, approve the connection, and the connector activates. No restart needed.
Step 4: Try it
Open a conversation in Claude and ask it to send something to noctua. For example:
“Read this article and summarise the key points as a fifteen-minute briefing, then send the briefing to noctua titled ‘Architecture Notes’ by ‘My Research’.”
Claude will call the convert_text_to_audio tool, you’ll see a confirmation, and within a minute the edition appears in your podcast feed.
Other AI apps
The same URL works in any MCP-compatible client. Claude on the web uses the same flow (Settings → Connectors → Add custom connector). Cursor, Continue, and most others expose an equivalent “add remote MCP server” setting — point it at https://api.noctua.uno/mcp and it’ll work.
A few practical notes
- Each conversion uses one credit, same as any other noctua reading.
- Content limit per submission is 500,000 characters — long enough for most documents, research summaries, and chat transcripts.
- The author and title are yours to set. Your AI will usually pick sensible ones; you can also tell it exactly what you want.
- Sign-in is one-time per device. Once you’ve authorised Claude Desktop, the token persists across sessions.
Where this is heading
The MCP server is the first step in a broader idea: noctua should sit at the end of every “I’ll get to that later” pipeline. Your inbox forwards into it. Your browser pastes into it. Your AI hands work off to it. The friction between worth reading and actually heard keeps shrinking.
If you build with MCP — or if you’ve wired up a clever automation around this — we’d love to hear about it.
Connect Claude, summarise something, and listen to it on your next walk.
— The noctua team