- Standardize invoke_agent pipeline span descriptions to use
invoke_agent (with optional functionId suffix) instead of Vercel SDK function names like generateText or generateObject. This aligns with how other AI integrations (e.g. LangGraph) name their agent spans.
- Unify all
.do* content generation span descriptions under a single generate_content prefix (e.g. generate_content mock-model-id) instead of using individual prefixes like generate_text, stream_text, generate_object, stream_object.
- Remove
addOriginToSpan helper and inline the setAttribute call directly.
Note
The pull request "fix(core): Standardize Vercel AI span descriptions to align with GenAI semantic conventions" was created by @RulaKhaled but did not reference an issue. Therefore this issue was created for better visibility in external tools like Linear.
invoke_agent(with optionalfunctionIdsuffix) instead of Vercel SDK function names likegenerateTextorgenerateObject. This aligns with how other AI integrations (e.g. LangGraph) name their agent spans..do*content generation span descriptions under a singlegenerate_contentprefix (e.g.generate_content mock-model-id) instead of using individual prefixes likegenerate_text,stream_text,generate_object,stream_object.addOriginToSpanhelper and inline thesetAttributecall directly.