Get Granular Usage
Get granular usage data with flexible grouping.
kind selects the billable usage domain:
traces(default): trace counts.langsmith_deployments: LangSmith Deployment metrics (nodes executed, agent runs, agent uptime). The three Deployment fields are populated andtracesis0.
trace_tier (only meaningful for kind=traces) optionally restricts
results to a single retention tier (longlived = extended retention,
shortlived = standard retention). When group_by=trace_tier, results
are split into one record per retention tier per time bucket.
workspace_ids filters results to the specified workspaces. Only
workspaces the user has read access to are included.
Authorizations
Query Parameters
Dimensions for grouping granular usage data.
user, api_key, project, workspace, trace_tier Which billable usage domain a granular-usage query targets.
traces: trace counts.langsmith_deployments: LangSmith Deployment metrics (nodes executed, agent runs, agent uptime).
Default is traces for backward compatibility — existing callers of
GET /granular-usage without a kind query param get the same
response shape they always did.
traces, langsmith_deployments longlived, shortlived 
