> ## Documentation Index
> Fetch the complete documentation index at: https://docs.langchain.com/llms.txt
> Use this file to discover all available pages before exploring further.

# LangSmith Observability

Traces are a series of steps that your application takes to go from input to output. Each of these individual steps is represented by a run. You can use [LangSmith](https://smith.langchain.com?utm_source=docs\&utm_medium=cta\&utm_campaign=langsmith-signup\&utm_content=oss-langgraph-observability) to visualize these execution steps. To use it, [enable tracing for your application](/langsmith/trace-with-langgraph). This enables you to do the following:

* [Debug a locally running application](/langsmith/observability-studio#debug-langsmith-traces).
* [Evaluate the application performance](/oss/javascript/langchain/test/evals).
* [Monitor the application](/langsmith/dashboards).

## Prerequisites

Before you begin, ensure you have the following:

* **A LangSmith account**: Sign up (for free) or log in at [smith.langchain.com](https://smith.langchain.com?utm_source=docs\&utm_medium=cta\&utm_campaign=langsmith-signup\&utm_content=oss-langgraph-observability).
* **A LangSmith API key**: Follow the [Create an API key](/langsmith/create-account-api-key) guide.

## Enable tracing

To enable tracing for your application, set the following environment variables:

```python theme={"theme":{"light":"catppuccin-latte","dark":"catppuccin-mocha"}}
export LANGSMITH_TRACING=true
export LANGSMITH_API_KEY=<your-api-key>
```

By default, the trace will be logged to the project with the name `default`. To configure a custom project name, see [Log to a project](#log-to-a-project).

For more information, see [Trace with LangGraph](/langsmith/trace-with-langgraph).

## Trace selectively

You may opt to trace specific invocations or parts of your application using LangSmith's `tracing_context` context manager:

```ts theme={"theme":{"light":"catppuccin-latte","dark":"catppuccin-mocha"}}
import { LangChainTracer } from "@langchain/core/tracers/tracer_langchain";

// This WILL be traced
const tracer = new LangChainTracer();
await agent.invoke(
  {
    messages: [{role: "user", content: "Send a test email to alice@example.com"}]
  },
  { callbacks: [tracer] }
);

// This will NOT be traced (if LANGSMITH_TRACING is not set)
await agent.invoke(
  {
    messages: [{role: "user", content: "Send another email"}]
  }
);
```

## Log to a project

<Accordion title="Statically">
  You can set a custom project name for your entire application by setting the `LANGSMITH_PROJECT` environment variable:

  ```bash theme={"theme":{"light":"catppuccin-latte","dark":"catppuccin-mocha"}}
  export LANGSMITH_PROJECT=my-agent-project
  ```
</Accordion>

<Accordion title="Dynamically">
  You can set the project name programmatically for specific operations:

  ```ts theme={"theme":{"light":"catppuccin-latte","dark":"catppuccin-mocha"}}
  import { LangChainTracer } from "@langchain/core/tracers/tracer_langchain";

  const tracer = new LangChainTracer({ projectName: "email-agent-test" });
  await agent.invoke(
    {
      messages: [{role: "user", content: "Send a test email to alice@example.com"}]
    },
    { callbacks: [tracer] }
  );
  ```
</Accordion>

## Add metadata to traces

You can annotate your traces with custom metadata and tags:

```ts theme={"theme":{"light":"catppuccin-latte","dark":"catppuccin-mocha"}}
import { LangChainTracer } from "@langchain/core/tracers/tracer_langchain";

const tracer = new LangChainTracer({ projectName: "email-agent-test" });
await agent.invoke(
  {
    messages: [{role: "user", content: "Send a test email to alice@example.com"}]
  },
  config: {
    tags: ["production", "email-assistant", "v1.0"],
    metadata: {
      userId: "user123",
      sessionId: "session456",
      environment: "production"
    }
  },
);

```

This custom metadata and tags will be attached to the trace in LangSmith.

<Tip>
  To learn more about how to use traces to debug, evaluate, and monitor your agents, see the [LangSmith documentation](/langsmith/home).
</Tip>

## Use anonymizers to prevent logging of sensitive data in traces

You may want to mask sensitive data to prevent it from being logged to LangSmith. You can create [anonymizers](/langsmith/mask-inputs-outputs#rule-based-masking-of-inputs-and-outputs) and apply them to
your graph using configuration. This example will redact anything matching the Social Security Number format XXX-XX-XXXX from traces sent to LangSmith.

```typescript TypeScript theme={"theme":{"light":"catppuccin-latte","dark":"catppuccin-mocha"}}
import { StateGraph } from "@langchain/langgraph";
import { LangChainTracer } from "@langchain/core/tracers/tracer_langchain";
import { StateAnnotation } from "./state.js";
import { createAnonymizer } from "langsmith/anonymizer"
import { Client } from "langsmith"

const anonymizer = createAnonymizer([
  // Matches SSNs
  { pattern: /\b\d{3}-?\d{2}-?\d{4}\b/, replace: "<ssn>" }
])

const langsmithClient = new Client({ anonymizer })
const tracer = new LangChainTracer({
  client: langsmithClient,
});

export const graph = new StateGraph(StateAnnotation)
  .compile()
  .withConfig({ callbacks: [tracer] });
```

***

<div className="source-links">
  <Callout icon="terminal-2">
    [Connect these docs](/use-these-docs) to Claude, VSCode, and more via MCP for real-time answers.
  </Callout>

  <Callout icon="edit">
    [Edit this page on GitHub](https://github.com/langchain-ai/docs/edit/main/src/oss/langgraph/observability.mdx) or [file an issue](https://github.com/langchain-ai/docs/issues/new/choose).
  </Callout>
</div>