Developer Ships Open-Source Java Library to Fill Agent Protocol's Missing Server Side
A developer has released Franca, an open-source Java 21 library published to Maven Central, designed to address a structural gap in how AI agent protocols are implemented. While client-side tools from vendors like OpenAI and Anthropic define conversation shape, tool declarations, and streaming semantics in detail, the server-side counterpart — where data, governance, and execution actually reside — remains largely unspecified. Currently, developers must rely on vendor-hosted platforms, protocol gateways like LiteLLM, or self-hosted inference servers that lack true agent semantics, leaving application-layer control in third-party hands. Franca terminates both OpenAI Responses and Anthropic Messages dialects within a developer's own service, connecting any OpenAI-compatible inference backend to standard endpoints via configurable adapters, drivers, and routes. The library integrates with Spring Boot 3 and aims to give application developers direct ownership of the server half of the agent stack.
This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)
Log in to join the discussion and vote.
Log in