> For the complete documentation index, see [llms.txt](https://help.nightfall.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://help.nightfall.ai/data-exfiltration-prevention/ai-agent-security/ai-governance/auditability-and-control/setup-and-installation/claude-cowork-opentelemetry.md).

# Claude Cowork (OpenTelemetry)

Claude Cowork can't run Nightfall hooks, so it's monitored through Anthropic's built-in OpenTelemetry (OTel) integration instead. You configure it once from the Anthropic admin console — not per device or per user — and it captures prompts, tool calls, file access, and session metadata across every Team / Enterprise seat.

OTel is **monitor-only**: it reports on activity after it happens and cannot block. For how it compares to hooks, see Hooks vs. OpenTelemetry.

{% hint style="info" %}
**Availability:** Requires a **Claude Team or Enterprise** plan, and the integration must be configured by an Anthropic **organization owner**.
{% endhint %}

**1. Copy your OTLP values from Nightfall**

In the Nightfall console, open **AI Agent Security → AI Agent Security Setup** and select the **OpenTelemetry** tab. Nightfall generates three values scoped to your organization — use **Copy all values** to grab them together:

| Field             | Value                                                                                  |
| ----------------- | -------------------------------------------------------------------------------------- |
| **OTLP Endpoint** | An HTTPS URL for your environment, e.g. `https://<your-nightfall-otel-host>/es/otel/…` |
| **OTLP Protocol** | `http/json`                                                                            |
| **OTLP Headers**  | `X-Nightfall-Company-Identifier=<company-id>,Authorization=Bearer <token>`             |

Copy each value exactly as shown — the endpoint host and the headers are specific to your organization and environment.

{% hint style="warning" %}
The **OTLP Headers** value contains a bearer token that authenticates your organization's telemetry. Treat it like a secret — copy it directly from the console into Anthropic, and don't paste it into tickets, chat, or screenshots.
{% endhint %}

**2. Paste them into the Anthropic console**

1. Sign in to `console.anthropic.com` as an organization owner. (You can use **Open Anthropic console** on the OpenTelemetry tab to jump straight there.)
2. Open **Organization Settings → Cowork → Observability**.
3. Paste the **endpoint**, **protocol**, and **headers** you copied from Nightfall.
4. Save.

**3. Verify the connection**

Back on the **OpenTelemetry** tab in Nightfall, the collector status card starts as **"OTEL collector not yet connected · Not configured."** Once Cowork sends its first telemetry — usually after the next Cowork session — the status flips to connected, and **Claude Cowork** appears as an active client on the **Devices** page.

{% hint style="info" %}
Connection status reflects a rolling **72-hour** activity window. If no Cowork telemetry is received for 72 hours, the collector shows as inactive again until the next session.
{% endhint %}

**What Cowork OTel captures**

Across every Team / Enterprise seat, with no per-device install:

* Prompts and session metadata
* Tool calls and file access
* Model name, token usage, and cost
* API errors and retries

Capturing full prompt text and tool input/output (rather than metadata only) depends on Anthropic's OTel logging options being enabled for your organization — see Hooks vs. OpenTelemetry and Anthropic's [Cowork monitoring docs](https://claude.com/docs/cowork/monitoring#events).


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://help.nightfall.ai/data-exfiltration-prevention/ai-agent-security/ai-governance/auditability-and-control/setup-and-installation/claude-cowork-opentelemetry.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.