From Natural Language to AI Agent Tool in 3 seconds!

Usage of tools is what separates an AI chatbot from an AI agent, and the more tools your AI agent has, the more capable it is.

You can create tools using Zapier, N8N, LangChain, Python, C#, Java, etc. However, all of these approaches requires you to understand either some difficult programming language, or some complex platform with "a bajillion" buttons and choices. With Magic Cloud you create tools using natural language, and more importantly; You create such tools in 3 seconds.

Unique tech

We have a unique platform, and we also have our own LLM. We refer to this LLM as the Hyperlambda Generator. It allows you to generate Hyperlambda code in 3 seconds, from natural language, implying 3 seconds after you've stated your intent, the tool is ready to be used. This allows you to phrase instructions to it such as follows.

• "Send an email to john@doe.com and attach the file '/report.pdf'"
• "Return 5 customers from ERP database sorted by registration_date"
• "Search DuckDuckGo for 'Thomas Hansen Hyperlambda' and return the first 5 matches"
• Etc ...

The point is that our LLM can do everything Hyperlambda can do, and it understands English. This arguably makes an entire suite of applications created to allow users to develop "tools" obsolete.

Evolutionary AI agents

In theory this allows us to deliver AI agents with zero tools, only having access to the "tool builder", and use this to dynamically build tools on the fly that somehow solves the user's problem.

Hence, even though such an AI agent starts out with zero tools, it still would have capabilities that no other AI agents in the world has, simply because of the "long tail problem", where "no two tools fits the same nail".

No two tools fits the same nail

Using a generic tool for a special problem is like trying to go grocery shopping with a Ferrari, and pick up the kids at school. The Ferrari is a really good car, but it wasn't created to go groceryu shopping, and it's not going to be comfortable for your kids either. Hence, use the right tool for the job!

However, in an evolutionary AI agent, you will always have access to the right tool, as long as you can somehow explain it using natural language. If you want to try this out, you can test it at our natural language API demo page.