# Open Agent Specification (Agent Spec) integration with LangGraph

**URL:** <https://forum.langchain.com/t/open-agent-specification-agent-spec-integration-with-langgraph/1775>\
**Category:** LangGraph\
**Tags:** product-feedback, python-help\
**Created:** [October 7, 2025, 10:08pm UTC](https://forum.langchain.com/t/open-agent-specification-agent-spec-integration-with-langgraph/1775 "2025-10-07T22:08:11Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![rhpatra](https://yyz1.discourse-cdn.com/flex007/user_avatar/forum.langchain.com/rhpatra/32/1203_2.png) [@rhpatra](https://forum.langchain.com/u/rhpatra)\
**Post date:** [October 7, 2025, 10:08pm UTC](https://forum.langchain.com/t/open-agent-specification-agent-spec-integration-with-langgraph/1775/1 "2025-10-07T22:08:11Z")

</div>

Oracle recently introduced the **Open Agent Specification (Agent Spec)** which is a framework-agnostic, declarative standard designed to make AI agents and workflows **portable, reusable, and executable** across any compatible framework.

[https://blogs.oracle.com/ai-and-datascience/post/introducing-open-agent-specification](https://blogs.oracle.com/ai-and-datascience/post/introducing-open-agent-specification)

A **LangGraph adapter** is already available, enabling initial interoperability - [agent-spec/adapters/langgraphagentspecadapter at main · oracle/agent-spec · GitHub](https://github.com/oracle/agent-spec/tree/main/adapters/langgraphagentspecadapter)

However, it would be highly valuable to see **native support within LangGraph** for **importing and exporting agents** using this specification. Such integration could pave the way for a truly **open Agent Registry** , much like how **ONNX standardized model exchange** across deep learning frameworks.