Introducing the Kangentic Blog
A new home for tutorials, cookbooks, and behind-the-scenes looks at building with AI coding agents.
Hey, Tyler here. I’ve been building Kangentic over the past several months, an open-source Kanban board for orchestrating AI coding agents like Claude Code, Codex, Gemini, Cursor, GitHub Copilot, Aider, and Warp. Along the way I’ve run into a lot of interesting engineering challenges, discovered workflows that genuinely changed how I ship code, and learned a ton about what works (and what doesn’t) when you’re coordinating multiple agents at once.
I’ve been dogfooding Kangentic as my daily driver the entire time, using it to build itself. It’s given me a front-row seat to what actually matters when you’re running multiple agents side by side all day. I wanted a place to capture all of that, so here’s the blog.
What I’ll be writing about
- Cookbooks and recipes - step-by-step patterns for multi-agent workflows. How to set up a Plan, Execute, and Review pipeline, how to run parallel agents across git worktrees, how to use MCP tools so agents can hand work to each other.
- Building fast with Kangentic - real projects, real timelines. How I use Kangentic to ship features, the workflows that stick, and the ones that don’t.
- AI agent deep-dives - the broader ecosystem is moving fast. I’ll cover what’s changing in agent orchestration, new CLI capabilities, and how to get the most out of your coding agents.
- Feature spotlights - when I ship something big, I’ll break down the design decisions, how it works, and how to get the most out of it.
Follow along
You can subscribe via RSS to get new posts in your feed reader, or just check back here. I’ll be posting regularly as I ship new features and discover new patterns.
If you haven’t tried Kangentic yet, get started in under a minute. It’s free and open source.