getsentry · sfanahata · Jul 22, 2026
diff --git a/docs/concepts/key-terms/tracing/index.mdx b/docs/concepts/key-terms/tracing/index.mdx
@@ -81,6 +81,10 @@ While everyone is sharing Lighthouse scores, we all know that the true determina

 Tracing can even give you the power to debug unfamiliar codebases. Although this benefit is not often discussed, a lot of time can be saved when debugging applications instrumented with Tracing, even if the codebase is completely new to you. Being able to trace actual user journeys throughout your application and see what happens across your entire stack contextualizes issues without even needing to recreate the issue locally. Read about other ways [tracing helps developers debug various issues](https://blog.sentry.io/everyone-needs-to-know-how-to-trace/).

+### Trace AI Agent Workflows
+
+If you're building with AI agents, LLMs, or MCP servers, Sentry's [Agent Tracing](/ai/observability/agents/) extends your existing tracing setup to capture the full picture — model calls, tool executions, token usage, and agent handoffs — all connected to the same traces as your HTTP requests, database queries, and application errors. Learn more about [setting up Agent Tracing](/ai/observability/agents/getting-started/).
+
 ## Learn More

 <PageGrid />
diff --git a/docs/platforms/dotnet/common/tracing/index.mdx b/docs/platforms/dotnet/common/tracing/index.mdx
@@ -64,6 +64,12 @@ Test out tracing by starting and finishing a transaction, which you _must_ do so

 While you're testing, set <PlatformIdentifier name="traces-sample-rate" /> to `1.0`, as that ensures that every transaction will be sent to Sentry. Once testing is complete, you may want to set a lower <PlatformIdentifier name="traces-sample-rate" /> value, or switch to using <PlatformIdentifier name="traces-sampler" /> to selectively sample and filter your transactions, based on contextual data.

+## Agent Tracing
+
+If you're building with AI agents or LLMs, you can extend your tracing setup to capture agent workflows, model calls, tool executions, and token usage. With tracing already enabled, the .NET SDK can automatically instrument AI agents built with `Microsoft.Extensions.AI`.
+
+<PlatformLink to="/tracing/instrumentation/ai-agents-module/">Set up Agent Tracing</PlatformLink> to get started.
+
 ## Next Steps

 <PageGrid />
diff --git a/docs/platforms/javascript/common/tracing/index.mdx b/docs/platforms/javascript/common/tracing/index.mdx
@@ -55,6 +55,24 @@ You can find more in-depth explanations and examples about sampling configuratio

 Sentry captures distributed traces consisting of transactions and spans, which measure individual services and individual operations within those services, respectively. Learn more about our model in [Distributed Tracing](/concepts/key-terms/tracing/distributed-tracing/).

+## Agent Tracing
+
+<PlatformCategorySection supported={["server", "serverless"]}>
+
+If you're building with AI agents or LLMs, you can extend your tracing setup to capture agent workflows, model calls, tool executions, and token usage. With tracing already enabled, the SDK can automatically instrument supported AI libraries and connect agent activity to the rest of your application traces.
+
+<PlatformLink to="/agent-tracing/">Set up Agent Tracing</PlatformLink> to get started.
+
+</PlatformCategorySection>
+
+<PlatformCategorySection supported={["browser"]}>
+
+If you're making AI or LLM calls from the browser, you can extend your tracing setup to capture model requests and token usage. With tracing already enabled, you can manually instrument AI spans and connect them to the rest of your application traces.
+
+<PlatformLink to="/agent-tracing-browser/">Set up Agent Tracing</PlatformLink> to get started.
+
+</PlatformCategorySection>
+
 ## Verify

 <PlatformSection supported={["javascript"]} notSupported={["javascript.cordova"]}>

diff --git a/docs/platforms/javascript/guides/cloudflare/features/workers-ai.mdx b/docs/platforms/javascript/guides/cloudflare/features/workers-ai.mdx
@@ -114,4 +114,4 @@ export const MyChatAgent = Sentry.instrumentDurableObjectWithSentry(

 To unset the conversation ID, pass `null` to `Sentry.setConversationId()`.

-Grouped AI calls show up in the <Link to="/ai/monitoring/conversations/">Conversations</Link> view, where you can inspect token usage, latency, and errors across an entire conversation.
+Grouped AI calls show up in the <Link to="/ai/observability/conversations/">Conversations</Link> view, where you can inspect token usage, latency, and errors across an entire conversation.
diff --git a/docs/platforms/php/common/tracing/index.mdx b/docs/platforms/php/common/tracing/index.mdx
@@ -46,6 +46,16 @@ Follow the instructions in the Relay docs to send a test event through Relay to
 Don't forget to update your `DSN` to point to your running Relay instance.
 After you set up Relay, you should see a dramatic improvement to the impact on your server.

+<PlatformSection supported={["php.laravel"]}>
+
+## Agent Tracing
+
+If you're building with AI agents or LLMs in Laravel, you can extend your tracing setup to capture agent workflows, model calls, tool executions, and token usage. With tracing already enabled, the Laravel SDK automatically instruments the [Laravel AI](https://laravel.com/docs/13.x/ai-sdk) package.
+
+<PlatformLink to="/agent-tracing/">Set up Agent Tracing</PlatformLink> to get started.
+
+</PlatformSection>
+
 ## Next Steps

 <PageGrid />
diff --git a/docs/platforms/python/tracing/index.mdx b/docs/platforms/python/tracing/index.mdx
@@ -35,6 +35,12 @@ If you’re adopting Tracing in a high-throughput environment, we recommend test

 Learn more about tracing <PlatformLink to="/configuration/options/#tracing-options">options</PlatformLink>, how to use the <PlatformLink to="/configuration/sampling/#setting-a-sampling-function">traces_sampler</PlatformLink> function, or how to <PlatformLink to="/configuration/sampling/#sampling-transaction-events">sample transactions</PlatformLink>.

+## Agent Tracing
+
+If you're building with AI agents or LLMs, you can extend your tracing setup to capture agent workflows, model calls, tool executions, and token usage. With tracing already enabled, the Python SDK automatically instruments supported libraries like OpenAI, Anthropic, LangChain, and others.
+
+<PlatformLink to="/tracing/instrumentation/custom-instrumentation/ai-agents-module/">Set up Agent Tracing</PlatformLink> to get started.
+
 ## Next Steps

 <PageGrid />
diff --git a/docs/platforms/react-native/common/tracing/index.mdx b/docs/platforms/react-native/common/tracing/index.mdx
@@ -37,6 +37,12 @@ Verify that tracing is working correctly by using our <PlatformLink to="/tracing

 While you're testing, set <PlatformIdentifier name="traces-sample-rate" /> to `1.0`, as that ensures that every transaction will be sent to Sentry. Once testing is complete, you may want to set a lower <PlatformIdentifier name="traces-sample-rate" /> value, or switch to using <PlatformIdentifier name="traces-sampler" /> to selectively sample and filter your transactions, based on contextual data.

+## Agent Tracing
+
+If your app makes AI or LLM calls, you can extend your tracing setup to capture model requests and token usage. With tracing already enabled, you can manually instrument AI spans and connect them to the rest of your application traces.
+
+<PlatformLink to="/agent-tracing/">Set up Agent Tracing</PlatformLink> to get started.
+
 ## Next Steps

 <PageGrid />
diff --git a/docs/platforms/ruby/common/tracing/index.mdx b/docs/platforms/ruby/common/tracing/index.mdx
@@ -37,6 +37,12 @@ Test out tracing by starting and finishing a transaction, which you _must_ do so

 While you're testing, set <PlatformIdentifier name="traces-sample-rate" /> to `1.0`, as that ensures that every transaction will be sent to Sentry. Once testing is complete, you may want to set a lower <PlatformIdentifier name="traces-sample-rate" /> value, or switch to using <PlatformIdentifier name="traces-sampler" /> to selectively sample and filter your transactions, based on contextual data.

+## Agent Tracing
+
+If you're building with AI agents or LLMs, you can extend your tracing setup to capture agent workflows, model calls, tool executions, and token usage. With tracing already enabled, you can use custom instrumentation to capture AI agent spans and connect them to the rest of your application traces.
+
+<PlatformLink to="/tracing/instrumentation/custom-instrumentation/ai-agents-module/">Set up Agent Tracing</PlatformLink> to get started.
+
 ## Next Steps

 <PageGrid />
diff --git a/docs/product/sentry-basics/performance-monitoring.mdx b/docs/product/sentry-basics/performance-monitoring.mdx
@@ -51,6 +51,7 @@ Performance data in Sentry isn't siloed. Every trace connects to:
 - **[Logs](/product/logs/)** — see what your application logged during a slow request
 - **[Session Replay](/product/session-replay/)** — watch the user experience during a performance issue
 - **[Profiling](/product/profiling/)** — drill from a slow span into the exact functions consuming CPU time
+- **[Agent Tracing](/ai/observability/agents/)** — trace AI agent workflows, LLM calls, and tool executions alongside your application traces
 - **[Seer](/product/ai-in-sentry/seer/)** — get AI-powered root cause analysis for performance regressions

 ## How It Works Under the Hood