Add AI telemetry configuration for Langfuse and Braintrust

content/configuration/ai.md

@@ -29,3 +29,37 @@ This is useful for compliance requirements where AI features must be completely
 ::callout{icon="material-symbols:info" color="info"}
 When `MCP_ENABLED` is set to `false`, the MCP server cannot be enabled through **Settings → AI → Model Context Protocol** in the admin interface, providing system administrators with complete control over AI integration features. See the [MCP Server](/guides/ai/mcp/installation) guide for more information.
 ::
+
+## Telemetry
+
+Send AI Assistant traces to an external observability platform for monitoring usage, performance, and cost.
+
+| Variable | Description | Default Value |
+| -------- | ----------- | ------------- |
+| `AI_TELEMETRY_ENABLED` | Enable OpenTelemetry-based tracing for AI Assistant requests. User IDs and Role ID will be included in traces. | `false` |
+| `AI_TELEMETRY_PROVIDER` | Telemetry provider to use. Supported values: `langfuse`, `braintrust`. | `langfuse` |
+| `AI_TELEMETRY_RECORD_IO` | Include full prompt inputs and response outputs in traces. | `false` |
+
+::callout{icon="material-symbols:warning" color="warning"}
+Enabling `AI_TELEMETRY_RECORD_IO` will send the full content of user messages and AI responses to your telemetry provider. Only enable this if your telemetry provider meets your data privacy requirements.
+::
+
+### Langfuse
+
+Set `AI_TELEMETRY_PROVIDER` to `langfuse` (default).
+
+| Variable | Description | Default Value |
+| -------- | ----------- | ------------- |
+| `LANGFUSE_SECRET_KEY` | Langfuse secret key. | — |
+| `LANGFUSE_PUBLIC_KEY` | Langfuse public key. | — |
+| `LANGFUSE_BASE_URL` | Langfuse API base URL. | `https://cloud.langfuse.com` |
+
+### Braintrust
+
+Set `AI_TELEMETRY_PROVIDER` to `braintrust`.
+
+| Variable | Description | Default Value |
+| -------- | ----------- | ------------- |
+| `BRAINTRUST_API_KEY` | Braintrust API key. | — |
+| `BRAINTRUST_PROJECT_NAME` | Braintrust project name for grouping traces. | — |
+| `BRAINTRUST_API_URL` | Braintrust API URL. Only needed for self-hosted instances. | — |

content/guides/11.ai/1.assistant/1.setup.md

@@ -326,6 +326,46 @@ For details on creating prompts with variables, see [MCP Prompts](/guides/ai/mcp
   - [Google AI](https://aistudio.google.com/)
 - Consider self-hosted models via [OpenAI-compatible providers](#openai-compatible-providers) for cost control
 
+## Telemetry
+
+Directus can send AI Assistant traces to [Langfuse](https://langfuse.com) or [Braintrust](https://braintrust.dev) for observability — monitor usage, latency, token costs, and full request/response content.
+
+Set the following environment variables to enable telemetry:
+
+```ini
+AI_TELEMETRY_ENABLED=true
+AI_TELEMETRY_PROVIDER=langfuse # or braintrust
+```
+
+### Langfuse
+
+```ini
+AI_TELEMETRY_PROVIDER=langfuse
+LANGFUSE_SECRET_KEY=sk-lf-...
+LANGFUSE_PUBLIC_KEY=pk-lf-...
+LANGFUSE_BASE_URL=https://cloud.langfuse.com # optional, default shown
+```
+
+![Langfuse trace dashboard showing AI Assistant requests](/img/ai-telemetry-langfuse.png)
+
+### Braintrust
+
+```ini
+AI_TELEMETRY_PROVIDER=braintrust
+BRAINTRUST_API_KEY=sk-...
+BRAINTRUST_PROJECT_NAME=my-project # optional
+```
+
+![Braintrust trace dashboard showing AI Assistant requests](/img/ai-telemetry-braintrust.png)
+
+Optionally enable `AI_TELEMETRY_RECORD_IO=true` to include full prompt inputs and AI responses in traces.
+
+::callout{icon="material-symbols:warning" color="warning"}
+Enabling `AI_TELEMETRY_RECORD_IO` will send the full content of user messages and AI responses to your telemetry provider. Only enable this if your provider meets your data privacy requirements.
+::
+
+See the [AI Configuration](/configuration/ai#telemetry) reference for all available variables.
+
 ## Next Steps
 
 ::card-group