Integrations / LangSmith
LangSmith
LLM runs with tokens and cost, per trace.
Why connect it
Same question for LangSmith projects: cost per trace and per run type, and which step dominates.
What you'll see
- Cost per trace over 14 days
- Dominant step per project
- Cache and model-share signals
What the credential can and cannot do
Where to get it: smith.langchain.com → Settings → API keys.
It grants: Read on runs in the workspace.
It cannot: GitDoctor never creates runs or datasets.
Whatever you connect lives in an encrypted cookie in your browser and nowhere else. Sign out and it's gone; revoke it at the provider and it stops working immediately.