An indentation-based, type-safe domain-specific language (DSL) for programming stateful multi-agent workflows, LLM routing, and structured execution graphs.
Thanks everyone for checking out SWAN L4!
L4 is a human-readable scripting language for building AI agents and workflows without the complexity of traditional programming. The goal is to make agent orchestration simple, predictable, and easy to read, even as workflows grow more sophisticated.
Some highlights:
* ðĶĒ Clean, natural-language syntax
* ðĪ Built-in support for AI reasoning, tools, and agent-to-agent communication
* ð Structured control flow (loops, conditionals, parallel execution, iteration)
* ð§ Explicit context and variable handling for reliable workflows
* ð Localized diagnostics and developer-friendly debugging
We're excited to hear your thoughts:
* What kinds of AI agents would you build with a language like this?
* Which features would make AI workflow development easier for you?
We'll be here all day to answer questions, collect feedback, and discuss ideas. Thanks for your support! ð
Report
No reviews yetBe the first to leave a review for L4: AI Agent Programming Language