Replies: 2 comments
-
|
The Langfuse LLM Playground supports tool calling and structured output schemas(1), but the available documentation doesn't specifically mention support for Gemini's built-in "Grounding with Google Search" or "URL Context" tools. The Playground currently supports defining custom tools with JSON schema definitions and testing prompts with tool calls(1). You can also pass additional provider-specific parameters by toggling "Additional Options" in the model selection dropdown(1). Since Gemini grounding tools are provider-specific features, you might try configuring them through the "Additional Options" section when using a Gemini model in the Playground. However, I don't have explicit documentation confirming this works for Gemini's grounding features. For tracing Gemini models outside the Playground, you can use the Google GenAI SDK with OpenTelemetry instrumentation to send traces to Langfuse(2), but this is for observability rather than Playground configuration. If you need more specific guidance on configuring Gemini grounding tools in the Playground, I'd recommend checking these potentially useful sources: 📚 Sources: Have another question? Just tag @inkeep. |
Beta Was this translation helpful? Give feedback.
-
|
Hi @mdbudnick, we recently added support for additional provider options for Gemini in the playground: https://github.com/orgs/langfuse/discussions/7928 |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
Describe your question
Gemini has built-in "Grounding with Google Search" and "URL Context" tools, but I don't know how to configure these in the LLM Playground.
Langfuse Cloud or Self-Hosted?
Langfuse Cloud
If Self-Hosted
No response
If Langfuse Cloud
No response
SDK and integration versions
No response
Pre-Submission Checklist
Beta Was this translation helpful? Give feedback.
All reactions