SuperSpec is our declarative DSL that makes agent building as simple as writing a specification.
Think of it as "Kubernetes for AI agents" - you describe what you want, and SuperOptiX builds the entire pipeline.
SuperSpec (pronounced /suห.pษr spษk/) is the context and agent engineering specification language for AI agents. It's designed to provide the just-right context to agents so they perform better.
Context engineering is the systematic approach to designing dynamic systems that deliver precisely the right information and tools in the optimal format, enabling LLMs to successfully accomplish their intended tasks.
When agents fail to perform reliably, the root cause is almost always insufficient or poorly structured context, unclear instructions, or missing tools that haven't been properly communicated to the model.
Agent engineering represents the next evolution of AI engineering. Rather than developing systems with static, hardcoded logic, engineers now design autonomous, goal-driven entities capable of using tools, accessing memory, engaging in reflective reasoning, and operating within safety constraints.
SuperSpec is declarative and strongly typed to ensure strong contracts between context and LLM output. This contract then converts into DSPy Signatures which validate the output even further.
Like Kubernetes DSL for declaring pods, deployments, and services, SuperSpec provides a Kubernetes-style declarative specification for AI agents. You declare what you want, not how to get it.
SuperSpec specifications are totally version controllable and context can be versioned, enabling Git-based agent management, rollback capabilities, A/B testing, and team collaboration.
Simple cloud-based demo agent for basic question-answering
Basic Structure
Execution Flow
Advanced Features