Google Released Agent Development Kit for Kotlin 1.0
The framework brings production-ready AI agent capabilities to the Kotlin Multiplatform ecosystem.
Updated on Sept. 20, 2026 in Artificial Intelligence

Live Poll
Would you prefer to build AI agents using native language frameworks like Kotlin?
Google has officially released the Agent Development Kit (ADK) for Kotlin 1.0, a production-ready framework designed for building AI agents across Android, JVM servers, and Kotlin Multiplatform applications. This release achieves feature parity with existing Python and Java ADK versions.
Why it matters
By leveraging Kotlin Multiplatform, this kit allows developers to share AI agent logic across diverse execution environments while maintaining native performance. It provides a standardized approach to complex agent behaviors like hierarchical systems and human-in-the-loop workflows.
The ADK 1.0 utilizes Kotlin Symbol Processing (KSP) to generate function schemas at compile time rather than relying on runtime reflection. This design choice aims to improve type safety and increase execution speed compared to reflection-heavy frameworks.
The players
A global technology company that builds large-scale AI infrastructure, cloud platforms, and the Kotlin development ecosystem.
The details
The framework enables hierarchical multi-agent systems and includes session management for pausing, serializing, and restoring agent states. Developers define capabilities through skills that load procedural knowledge dynamically from files. For sensitive operations, the ADK supports human-in-the-loop workflows, requiring explicit user confirmation. Integration is provided for on-device inference via LiteRT-LM and ML Kit, alongside cloud functionality through Firebase AI Logic.
Timeline
September 20, 2026: Google released the Agent Development Kit for Kotlin 1.0.
The Tech Race
The release positions the Kotlin ecosystem to compete with established agent frameworks in the Python and Java environments. It extends the utility of Kotlin Multiplatform by standardizing how agents interact with both on-device and cloud-based inference stacks.
Developers building for Android or JVM backends can now implement complex AI agents with shared, type-safe logic. The framework is immediately accessible as an open-source tool for integration into existing production pipelines.
The takeaway
This release centralizes agent orchestration logic within the Kotlin ecosystem, reducing the need for language-specific implementations. Developers should watch for future updates regarding integration with emerging model inference benchmarks.
Further reading
For more on evolving frameworks in this space, visit Artificial Intelligence.
More information
Access the source code and documentation at the ADK for Kotlin GitHub repository.
Live Poll
Would you prefer to build AI agents using native language frameworks like Kotlin?









