apache/flink-agents

This system is a framework for building and executing AI agents within Apache Flink streaming pipelines. It provides a unified API for defining agent workflows, managing short-term and long-term memory, and integrating with various LLM providers and vector stores. The architecture supports cross-language interoperability between Java and Python, allowing developers to construct complex, stateful agent applications that leverage external tools and context retrieval.

This period marks the initial release of the Flink Agents framework, introducing a comprehensive Python and Java SDK for building and executing agent-based applications. The work focused on establishing the core runtime, plan, and API layers, while enabling cross-language interoperability between Python and Java components. Extensive testing and tooling were added to support the new functionality and ensure compatibility across different Flink versions.

70 changes

This period focused on expanding the Flink Agents framework with diverse AI model integrations, including OpenAI, Azure OpenAI, Google Gemini, and AWS Bedrock, alongside support for the Model Context Protocol. The work also introduced new agent capabilities such as Bash tool execution, declarative YAML configuration, and skill management, while improving the Python build process and adding comprehensive test coverage.

20 changes

apache/flink-agents was measured the same way every project in this corpus was: the same rubric, at a pinned commit, with the result published in full. Point the surveyor at a repository you know and see whether you agree with it.