In my previous blog post, I introduced the dynamic multi-agent app: an orchestrator that delegates to specialist agents, all configured at runtime through an admin UI: https://community.sap.com/t5/technology-blog-posts-by-members/agentic-ai-on-btp-dynamic-multi-agent-on-demand-with-pydantic-ai/ba-p/14388032

What I didn’t cover in that post is the layer every agent depends on: how a model name like gpt-4o or anthropic–claude-4-sonnet actually becomes a working LLM connection to SAP AI Core. Pydantic AI has never heard of SAP AI Core and SAP AI Core doesn’t know Pydantic AI. Nevertheless, with surprisingly little glue code they work nicely together, for OpenAI models and Claude models alike.

In this blog post, I’ll walk you through how that integration works: the problem, the integration logic for both model families and the lessons I learned along the way. All the code is in the btp-dynamic-multiagent-app repo, almost entirely in agents/shared.py.

​ In my previous blog post, I introduced the dynamic multi-agent app: an orchestrator that delegates to specialist agents, all configured at runtime through an admin UI: https://community.sap.com/t5/technology-blog-posts-by-members/agentic-ai-on-btp-dynamic-multi-agent-on-demand-with-pydantic-ai/ba-p/14388032What I didn’t cover in that post is the layer every agent depends on: how a model name like gpt-4o or anthropic–claude-4-sonnet actually becomes a working LLM connection to SAP AI Core. Pydantic AI has never heard of SAP AI Core and SAP AI Core doesn’t know Pydantic AI. Nevertheless, with surprisingly little glue code they work nicely together, for OpenAI models and Claude models alike.In this blog post, I’ll walk you through how that integration works: the problem, the integration logic for both model families and the lessons I learned along the way. All the code is in the btp-dynamic-multiagent-app repo, almost entirely in agents/shared.py.   Read More Technology Blog Posts by Members articles 

#SAP

#SAPTechnologyblog

In my previous blog post, I introduced the dynamic multi-agent app: an orchestrator that delegates to specialist agents, all configured at runtime through an admin UI: https://community.sap.com/t5/technology-blog-posts-by-members/agentic-ai-on-btp-dynamic-multi-agent-on-demand-with-pydantic-ai/ba-p/14388032

What I didn’t cover in that post is the layer every agent depends on: how a model name like gpt-4o or anthropic–claude-4-sonnet actually becomes a working LLM connection to SAP AI Core. Pydantic AI has never heard of SAP AI Core and SAP AI Core doesn’t know Pydantic AI. Nevertheless, with surprisingly little glue code they work nicely together, for OpenAI models and Claude models alike.

In this blog post, I’ll walk you through how that integration works: the problem, the integration logic for both model families and the lessons I learned along the way. All the code is in the btp-dynamic-multiagent-app repo, almost entirely in agents/shared.py.

​ In my previous blog post, I introduced the dynamic multi-agent app: an orchestrator that delegates to specialist agents, all configured at runtime through an admin UI: https://community.sap.com/t5/technology-blog-posts-by-members/agentic-ai-on-btp-dynamic-multi-agent-on-demand-with-pydantic-ai/ba-p/14388032What I didn’t cover in that post is the layer every agent depends on: how a model name like gpt-4o or anthropic–claude-4-sonnet actually becomes a working LLM connection to SAP AI Core. Pydantic AI has never heard of SAP AI Core and SAP AI Core doesn’t know Pydantic AI. Nevertheless, with surprisingly little glue code they work nicely together, for OpenAI models and Claude models alike.In this blog post, I’ll walk you through how that integration works: the problem, the integration logic for both model families and the lessons I learned along the way. All the code is in the btp-dynamic-multiagent-app repo, almost entirely in agents/shared.py.   Read More Technology Blog Posts by Members articles 

#SAP

#SAPTechnologyblog