🧪 Meet the Content Agent

We’re excited to introduce a powerful new feature inside Magnity: the Content Agent.

The Content Agent gives you a flexible space to experiment, strategize, and create — using your own content as the foundation. Whether you're working with web pages, documents, or videos, you can now ask your Content Agent anything, test messaging ideas, build automation flows, campaign briefs or translate content on the fly. Basically anything you can think of asking your content about.

🧩 What You Can Do

Just drag and drop any number of webpages, PDFs, or videos into the Content Agent canvas or into the chat. The Content Agent will instantly parse and understand your content — and you’re ready to go. Either you can modify the content or mix the content by dragging it into the chat or you can chat with the content only, by dragging it into the chat. You can even drag any content element or group of elements into the chat at any time, to add context to your conversartion.

From there, you can:

• 🤖 Ask questions and get insights from your own material, like: "Can you build an FAQ of 20 questions based on these videos?"
• ✉️Crearte campaign briefs, plans and even analyze your current content and identify gaps in messaging.
• 📈 Build plans for onboarding, engagement, nurturing, ABM or any other type of flow you can image. You can get anything from flows with funnel phases, start/end states, relevant URLs, and email breakdowns. You think it, we build it.
• 🌐 Translate any asset into any supported language in seconds and export it as a PDF.
• 🔍 Test prompts, messaging angles, and strategies — instantly
• 💬Drag content into the chat after having chatted about a topic for a while to add context to your conversation.

Magnity currently supports dragging in content parsing up to 60.000 tokens via ChatGPT, but we can expand that if you need to. We want to start off leaving a little space for conversation since the upper limit is 128.000 tokens. That’s enough to work with hundreds of pages at once — and we’re just getting started.

If you are curious about how tokens are counted, test it out here: Tokenizer - OpenAI API

Here are some examples of simple tasks:

1. “Generate a one-pager PDF outline from this video transcript.”
2. “Turn these product pages into comparison charts and FAQs for a landing page.”
3. Write a LinkedIn carousel outline with 5 slides highlighting insights from these pages.”
4. “Map these 20 pages to decision-journey phases (Aware, Consider, Decide).”
5. “Cluster these docs into themes with recommended campaign angles.”
6. “Extract pain points and value drivers for "persona a" based on these sources.”

We are going to overflood the new Content Agent Quick Prompt library with options and help, so you can get inspired on how to apply AI to your content library.

And yes, we will add the new drag content into chat for the emails, landing pages and SoMe posts as well, but it will be in the coming 1-2 weeks.

🔧 Pro Tips

• The Content Agent has rules consisiting of Writing rules and Tone of Voice guidelines. You can use these to form your responses.
• You can manully copy paste any text into the canvas that you want to work with.
• You can now add a tile to the chat by clicking a small button in the lower right corner of it.
• You can mark all the visible tiles in the content library and drag them all into the canvas or chat in one go.
• When you add one or more tiles to the chat, you can view the source content by clicking the small blob above the chat box. Here you can also remove them all or one by one.

🛠 Bug Fixes & Improvements

• Fixed an issue where the subject line would default to the master language if it contained a '%' sign when exporting to Marketo.
• Fixed an issue that caused the headline and image text in a 4 element lwebpage module to ignore the tag in the prompt, thereby doing generic headlines
• Cleaned up some minor UI issues.

🔍 Known Issues

• PDF documents has a bug that prevents you from adding them with all text to the canvas. We are fixing this within one week. We could have waited on releasing the rest, but thought we wanted you to try it out while we did some small fixes.
• If you want to add a document to the content library, do it from the email composer and not the content agent. When you add it ffrom the content agent, it does not show up in the library. We are fixcing this asap.
• Right now all responses are in plain text. We have discovered a bug in our Chat API that cause long HTML strings to deform, and are replacing the API method within the nexct week as well, so you can get well formatted responses.

🎯 Why This Matters

The Content Agent is a key step in evolving Magnity into a fully autonomous AI co-pilot — not just for content generation, but for decision-making, strategy, and campaign orchestration. Whether you're testing ideas or building production-ready flows, the Playground gives you creative control with AI-powered speed. We are researching how to expand the content agent to understsnd your entire information library from the get go, so you dont have to drsg anything in, but can just have an open conversation. Stay tuned and reach out to us, if you are an admin and we can tell you more.

Questions or Feedback?

Questions or Concerns?
Our support team is here to help:
✉️ ulsa@amagnity.ai | 📞 +45 29 35 25 78

Thank you for using Magnity!

— The Magnity Development Team