[Luca Forstner (lforst)]

• Add separate controls for chain inputs, chain outputs, model inputs, model outputs, and token usage in LangChain and LangGraph traces. Capture remains enabled by default.
• Skip disabled payloads before copying them, while keeping the trace structure, timing, errors, and metadata. This avoids copying large graph state or recording model data already captured by provider instrumentation.
• Support these options in manual callback handlers and automatic instrumentation:

  import { configureInstrumentation } from "braintrust/instrumentation";
  
  // Configure capture before SDK initialization or loading the tracing hook.
  configureInstrumentation({
    langchain: {
      captureChainInputs: false,
      captureChainOutputs: false,
      captureModelInputs: false,
      captureModelOutputs: false,
      captureModelTokenUsage: false,
    },
  });

[chatgpt-codex-connector]

Codex Review Summary: ✅ Completed

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-11-09T10:39:08.480011Z	3c6689a	PR opened

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "Codex (@codex) review" or "Codex (@codex) security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.