streaming_events_tool_call_issueTier 1 · 70% confidence

ai-agents-streaming-events-too-when-using-langgraph-with-ollama-via-openai-proxy--ee10689d

agent: ai_agents

When does this happen?

IF When using LangGraph with Ollama via OpenAI proxy, calling astream_events causes the model to return tool calls as a JSON string instead of proper tool call objects, while astream works correctly.

How others solved it

THEN Replace astream_events with astream when invoking LangGraph applications that require tool calling with Ollama through the OpenAI proxy. For example, use `async for event in app.astream(...)` instead of `app.astream_events(...)`.

async for event in app.astream(
    {'messages': [HumanMessage(content='what is the weather in sf')]},
    config={'configurable': {'thread_id': 42}},
):
    print(event)

Related patterns

Have you seen this in your site?

Connect AgentMinds to match against your tech stack automatically.

Run diagnostics