Currently, we have two SDKs, one for Python and another one for Node/TypeScript. Both SDKs in essence consist of two components:

• OpenTelemetry-based tracing, thanks to OpenLLMetry.
• Wrappers and convenience around Laminar's APIs, e.g. pipeline run requests or submit evaluation results.

With some demand for Go SDK, we would like to understand two things:

1. Do you use LLMs in Go? If so, how? What libraries, and how do you make calls?
2. Would the rest of Laminar's functionality (for now, primarily pipeline runs) be useful as a standalone SDK in Go, or would you rather send a plain API request from Go?

For the first part, OpenLLMetry did start experimenting with Go before https://github.com/traceloop/go-openllmetry, but it didn't seem to take off. We could do something similar though I'd approach it slightly differently. We are looking for suggestions though, and if this grows into a package of OpenTelemetry auto-instrumentations we will include it in Laminar after contributing it back to OpenLLMetry.