# MCP Integration

Dstl8 exposes an MCP (Model Context Protocol) server for each organization, giving AI assistants direct access to your logs, incidents, and observability tools.

### Get your MCP credentials

#### 1. Find your MCP endpoint

Each organization has a unique MCP endpoint:

```
https://<org_id>.app.dstl8.ai/mcp
```

Find your full URL in **Org** **Settings → Organizational Info.**  The `org_id` is also surfaced in the product sign-in URL for the Dstl8 UI.

#### 2. Generate an API token

1. Go to **Org Settings → API Tokens**
2. Click on **Create API Token**
3. Give it a descriptive name (e.g. "Claude Code - Jon")
4. Copy the token — it won't be shown again

Your token will look like `dstl8_xxxxxxxx_xxxxxxxxxxxxxxxxxxxxxxxx`.

### Available tools

Once connected, your AI assistant has access to:

* **Log querying** — search, filter, and analyze log samples and patterns
* **Incident management** — create, update, and list incidents
* **Anomaly detection** — surface anomalies across your services
* **Sentiment and severity analysis** — understand log health at a glance
* **Knowledge graph** — explore entities and relationships across your system

### Client setup guides

Choose your AI assistant:

* [Claude Code](/controltheory-documentation/dstl8-docs/mcp-integration/claude-code.md)
* [Claude Desktop](/controltheory-documentation/dstl8-docs/mcp-integration/claude-desktop.md)
* [Codex](/controltheory-documentation/dstl8-docs/mcp-integration/codex.md)
* [ChatGPT](/controltheory-documentation/dstl8-docs/mcp-integration/chatgpt.md)
* [Cursor](/controltheory-documentation/dstl8-docs/mcp-integration/cursor.md)

### Example prompts

Once connected, try asking:

**Situational awareness**

* "What are my active incidents?"
* "What's going on in production?"
* "Give me a health pulse across all services"
* "Any active incidents in staging?"

**Targeted investigation**

* "Why is the auth service throwing 500s?"
* "Is this error happening in staging too, or just prod?"
* "Show me recurring patterns in the payments service"

**Deploy verification**

* "Did my last deploy fix the connection timeout issue?"
* "Compare error rates before and after the 2pm deploy"
* "Is staging converging with production after the hotfix?"

**Pre-coding context**

* "What should I know about the billing service before I make changes?"
* "Any known incidents or patterns related to the webhook handler?"
* "Show me recent anomalies in the service I'm about to refactor"

**Cross-environment correlation**

* "Is the error I'm seeing locally also showing up in prod?"
* "Compare my local logs against staging for the same service"
* "Is this safe to promote to production?"


---

# Agent Instructions: Querying This Documentation

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

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.controltheory.com/controltheory-documentation/dstl8-docs/mcp-integration.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
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.