Python code interpreter sessions are generally available for Microsoft’s Azure Container Apps serverless platform for running applications and microservices. The service is also offering custom dynamic sessions, with JavaScript code interpreter capabilities in public preview.
Microsoft announced the general availability of the Python interpreter and custom dynamic sessions on November 19. Dynamic sessions provide fast access to built-in Python code interpreter sandboxes without the need to manage containers. To build advanced AI agents or copilots, large language models (LLMs) often are paired with a code interpreter. The interpreter extends the agent’s ability to perform complex tasks such as solving mathematical and reasoning problems or analyzing data, Microsoft said.
A dynamic session includes a Python code interpreter, which provides sandboxes to execute LLM-generated code in production. Developers can leverage dynamic sessions in LangChain, LlamaIndex, and Semantic Kernel agents with a few lines of code, according to Microsoft. In other applications needing to run untrusted Python code, developers can integrate with Python code interpreter sessions via an HTTP API.
In addition to the built-in Python code interpreter, dynamic sessions can run any custom container. Custom container dynamic sessions now support managed identity. These custom container sessions can be used to create a code container for specific needs, such as preinstalling dependencies or supporting a different language.
JavaScript is also supported in Azure Container Apps dynamic sessions. JavaScript code interpreter sessions can be used in applications to run untrusted code on the Node.js runtime. Developers can get started by creating a Node.js session portal in the Azure portal.