Embabel Agent Framework Reaches 1.0, Enabling Java and Kotlin AI Agents
A new open-source framework built on Spring AI lets developers define agents as typed domain objects with state machines and multi-model support.
Embabel has reached its 1.0 release, providing a framework for building AI agents in Java and Kotlin. The framework allows developers to define agents as typed domain objects, leveraging Spring AI for integration with multiple model providers. Embabel combines planning capabilities with predefined state machines, offering flexibility for designing agent workflows.
The release marks a milestone for Java and Kotlin developers seeking to incorporate AI agents into their applications. By building on Spring AI, Embabel taps into a well-established ecosystem, potentially lowering the barrier for enterprise adoption of agentic AI.
Why it matters
This release signals the maturation of AI agent frameworks beyond Python-centric ecosystems. As enterprises increasingly seek to integrate AI agents into existing Java and Kotlin codebases, frameworks like Embabel provide a path that aligns with familiar development patterns. The combination of typed domain objects and state machines offers a structured approach to agent design, which can improve reliability and maintainability in production systems.
This release signals the maturation of AI agent frameworks beyond Python-centric ecosystems.
Receive input
User or system sends a request to the agent
Plan action
Agent uses a model to determine the next steps
Execute state machine
Agent transitions through predefined states (e.g., fetch data, call API)
Generate response
What you can learn from this
- Typed domain objects for agents: Instead of treating agents as black boxes, Embabel lets you define them as typed domain objects. This means you can enforce compile-time checks on agent properties and behaviors, reducing runtime errors. Practice by modeling a simple customer support agent as a class with typed fields for intent, context, and response format.
- State machines for workflow control: Predefined state machines give you explicit control over agent transitions, such as moving from 'idle' to 'processing' to 'responding'. This is more predictable than purely generative flows. Try implementing a finite state machine for a multi-step task like order processing, where each state triggers a specific action.
- Multi-model provider support: By using Spring AI, Embabel abstracts away differences between model providers (e.g., OpenAI, Anthropic, local models). This lets you swap models without rewriting agent logic. Experiment by building a simple agent that queries two different models for the same prompt and compares responses.
- Planning combined with state machines: Agents can plan a sequence of actions and then execute them within a state machine, blending flexibility with structure. For a learning project, create an agent that plans a travel itinerary (planning phase) and then executes bookings step-by-step (state machine phase).
- Java/Kotlin for AI agents: This framework shows that AI agent development is not limited to Python. If you're a Java or Kotlin developer, you can leverage your existing skills to build agents. Start by integrating Embabel into a Spring Boot application and creating a simple agent that responds to user queries with a fixed workflow.
We teach this
Sources
Our reporting is an original summary; full coverage is at the links above.
Don't just read about it — build it.
Square 1 teaches the skills behind the headlines, with every line of your work graded by AI. Find your starting point in 3 minutes.
Get your free skill reportMore in Cloud
AWS Builder Center Offers Free Sandbox Environments for Workshops, Removing Cost Barrier for Learners
Read →Enterprise cloud infrastructure spending hits $143 billion per quarter, growth accelerating
Read →Cloudflare Unveils Meerkat: A Strongly Consistent Global Coordination Service Based on QuePaxa
Read →