Camel 4.22 brings two features that change how Camel integrates with AI: a unified tool abstraction (camel-ai-tool) and a built-in MCP server (camel-mcp-server). Together, they let you turn any Camel route into an AI tool that works across frameworks and protocols. This post covers both features. The Problem: Fragmented Tool Definitions Before 4.22, defining AI tools in Camel meant picking a framework and committing to it. Want to expose a route as a tool for LangChain4j?

Continue reading ❯

AICAMEL

Camel 4.21.0 shipped on 1 July with 32 security fixes. 4.18.3 followed on 3 July with 34, and 4.14.8 on 4 July with 27. All three supported release lines were patched inside four days, and every finding got a signed public advisory. A batch of 32 gets noticed, usually for the wrong reason. So I want to write about the process that produced it rather than walk through the list. If you are running Camel in production, the number of advisories in a release tells you very little.

Continue reading ❯

SECURITY

Apache Camel has always excelled at connecting systems, but observing what happens inside your running integrations has traditionally meant switching between CLI commands, log files, or firing up a web console. With Camel 4.21, we are introducing something new. Camel TUI is an interactive terminal dashboard that gives you a live view of your Camel integrations as they run on your local machine. Built on top of the Camel CLI, it renders real-time data with a modern terminal UI — mouse support, keyboard navigation, Unicode graphics, and even pixel-rendered route diagrams.

Continue reading ❯

FEATURES