What's new in n8n 1.100.1 · Content Academy

What's new in n8n 1.100.1

-AI Agent Tool node:

The AI Agent Tool node enables a simplified pattern for multi-agent orchestration that can be run in a single execution and stays entirely on one canvas. You can now connect multiple AI Agent Tool nodes to a primary AI Agent node, allowing it to supervise and delegate work across other specialized agents.

This setup is especially useful for building complex systems that function like real-world teams, where a lead agent assigns parts of a task to specialists. You can even add multiple layers of agents directing other agents, just like you would have in a multi-tiered organizational structure. It also helps with prompt management by letting you split long, complex instructions into smaller, focused tasks across multiple agents.

AI Agent Tool nodes are a good choice when you want the agent interaction to be completed in a single execution or prefer to manage and debug everything from a single canvas. Learn more.

-Built-in Metrics for AI Evaluations:

Evaluating AI is a best practice any AI solution, and a must if reliability and predictability are business-critical. Setting up evaluations for AI workflows is now easier than ever with built-in metrics.

Apply these metrics to review AI responses and assign scores based correctness, helpfulness, string similarity, categorization, or tools used. Run regular evaluations and review scores over time as a way to monitor your AI workflow’s performance. Compare results across different models to help guide model selection, or run evaluations before and after a prompt change to support data-driven, iterative building.

Learn more about built-in metrics or explore best practices for evaluating AI workflows OnDemand.

-Model Selector node:

The Model Selector node gives you more control when working with multiple LLMs in your workflows. Use it to direct which connected model should handle a given input, based on conditions you define. This node is especially useful in evaluation or production scenarios where routing logic between models needs to adapt based on performance, cost, availability, or dataset-specific needs.

To use, connect multiple LLMs to the Model Selector node, then configure routing conditions in the node’s settings. Learn more

-Convert to sub-workflow:

Sub-workflows allow you to take a more modular approach to building, breaking up big workflows into smaller, manageable parts that are easier to reuse, test, understand, and explain. Until this release, creating sub-workflows required copying and pasting nodes manually, setting up a new workflow from scratch, and reconnecting everything by hand. Convert to sub-workflow allows you to simplify this process into a single action, so you can spend more time building and less time restructuring.

To use:

Highlight the nodes you want to convert to a sub-workflow.
Right-click and select Convert to sub-workflow from the context menu (or Alt+X)

This will open a new tab containing the selected nodes, preserve all node parameters, and replace the selected nodes in the original workflow with a Call My Sub-workflow node.

Learn more about sub-workflows.

For all the details please open the image, it's a long screenshot.

4 comments