Agent Builder
Agent Builder is part of OpenAI’s tooling for constructing agentic applications, systems that use large language models to perform multi-step tasks autonomously, with governance, tool integration, memory, orchestration, and observability baked in. The platform offers a composable set of primitives—models, tools, memory/state, guardrails, and workflow orchestration- that developers assemble into agents capable of deciding when to call a tool, when to act, and when to halt and hand off control. OpenAI provides a new Responses API that combines chat capabilities with built-in tool use, along with an Agents SDK (Python, JS/TS) that abstracts the control loop, supports guardrail enforcement (validations on inputs/outputs), handoffs between agents, session management, and tracing of agent executions. Agents can be augmented with built-in tools like web search, file search, or computer use, or custom function-calling tools.
Learn more
Llama Stack
Llama Stack is a modular framework designed to streamline the development of applications powered by Meta's Llama language models. It offers a client-server architecture with flexible configurations, allowing developers to mix and match various providers for components such as inference, memory, agents, telemetry, and evaluations. The framework includes pre-configured distributions tailored for different deployment scenarios, enabling seamless transitions from local development to production environments. Developers can interact with the Llama Stack server using client SDKs available in multiple programming languages, including Python, Node.js, Swift, and Kotlin. Comprehensive documentation and example applications are provided to assist users in building and deploying Llama-based applications efficiently.
Learn more
Mistral Agents API
Mistral AI has introduced its Agents API, a significant advancement aimed at enhancing the capabilities of AI by addressing the limitations of traditional language models in performing actions and maintaining context. This new API integrates Mistral's powerful language models with several key features, built-in connectors for code execution, web search, image generation, and Model Context Protocol (MCP) tools; persistent memory across conversations; and agentic orchestration capabilities. The Agents API complements Mistral's Chat Completion API by providing a dedicated framework that simplifies the implementation of agentic use cases, serving as the backbone of enterprise-grade agentic platforms. It enables developers to build AI agents capable of handling complex tasks, maintaining context, and coordinating multiple actions, thereby making AI more practical and impactful for enterprises.
Learn more
Javalin
You never extend classes and you rarely implement interfaces. Javalin is just a few thousand lines of code on top of Jetty, and its performance is equivalent to raw Jetty code. Due to its size, it's very easy to reason about the source code. Other Java and Kotlin web frameworks usually offer one version for each language. Javalin is being made with inter-operability in mind, apps are built the same way in both Java and Kotlin. Javalin runs on top of Jetty, one of the most used and stable web servers on the JVM. You can configure the Jetty server fully, including SSL and HTTP3 and everything else that Jetty offers. Many lightweight Java and Kotlin web frameworks don't support OpenAPI, but Javalin does.
Learn more
