diff --git a/package.json b/package.json
index 77031c805f..c5481b57f3 100644
--- a/package.json
+++ b/package.json
@@ -43,6 +43,7 @@
     "packages/sdk/server-ai/examples/openai",
     "packages/sdk/server-ai/examples/tracked-chat",
     "packages/sdk/server-ai/examples/chat-observability",
+    "packages/sdk/server-ai/examples/openai-observability",
     "packages/sdk/server-ai/examples/vercel-ai",
     "packages/telemetry/browser-telemetry",
     "packages/sdk/combined-browser",
diff --git a/packages/sdk/server-ai/examples/openai-observability/.env.example b/packages/sdk/server-ai/examples/openai-observability/.env.example
new file mode 100644
index 0000000000..5863740ada
--- /dev/null
+++ b/packages/sdk/server-ai/examples/openai-observability/.env.example
@@ -0,0 +1,12 @@
+# LaunchDarkly SDK Key (required)
+LAUNCHDARKLY_SDK_KEY=your-launchdarkly-sdk-key-here
+
+# AI Config key (optional, defaults to 'sample-ai-config')
+LAUNCHDARKLY_AI_CONFIG_KEY=sample-ai-config
+
+# Observability service identification (optional)
+SERVICE_NAME=hello-js-openai-observability
+SERVICE_VERSION=1.0.0
+
+# OpenAI API Key (required)
+OPENAI_API_KEY=your-openai-api-key-here
diff --git a/packages/sdk/server-ai/examples/openai-observability/README.md b/packages/sdk/server-ai/examples/openai-observability/README.md
new file mode 100644
index 0000000000..2e43d77f58
--- /dev/null
+++ b/packages/sdk/server-ai/examples/openai-observability/README.md
@@ -0,0 +1,67 @@
+# Provider-Specific Observability Example (OpenAI)
+
+This example shows how to use the LaunchDarkly observability plugin when calling an AI provider directly — without the higher-level `createChat` abstraction. It uses OpenAI as the provider, but the same pattern applies to any provider (Bedrock, Anthropic, Vercel AI SDK, etc.).
+
+## How it works
+
+1. **Initialize the LaunchDarkly client** with the `Observability` plugin — this enables automatic capture of SDK operations, flag evaluations, errors, logs, and distributed traces.
+2. **Get the AI Config** via `completionConfig()` — this returns the model, messages, and parameters configured in LaunchDarkly, along with a `tracker` for reporting metrics.
+3. **Call your provider directly** and wrap it with the tracker — the tracker records latency, token usage, and success/error status.
+
+The tracker provides several methods depending on your provider. This example uses `trackMetricsOf` with the LaunchDarkly OpenAI provider's `getAIMetricsFromResponse` extractor:
+
+| Method | Provider |
+|--------|----------|
+| `tracker.trackMetricsOf(OpenAIProvider.getAIMetricsFromResponse, fn)` | OpenAI (recommended) |
+| `tracker.trackBedrockConverseMetrics(response)` | AWS Bedrock |
+| `tracker.trackVercelAISDKGenerateTextMetrics(fn)` | Vercel AI SDK |
+| `tracker.trackMetricsOf(extractor, fn)` | Any provider (custom extractor) |
+
+## Prerequisites
+
+1. A LaunchDarkly account and SDK key
+2. Node.js 16 or later
+3. Node server SDK v9.10 or later (required for the observability plugin)
+4. An OpenAI API key
+
+## Setup
+
+1. Install dependencies:
+
+   ```bash
+   yarn install
+   ```
+
+2. Set up environment variables:
+
+   ```bash
+   cp .env.example .env
+   ```
+
+   Edit `.env` and add your keys.
+
+3. Create an AI Config in LaunchDarkly (e.g. key `sample-ai-config`) with a completion-enabled variation and the model you want to use.
+
+## Running the Example
+
+```bash
+yarn start
+```
+
+This will:
+- Initialize the LaunchDarkly client with the observability plugin
+- Retrieve the AI Config (model, messages, parameters) from LaunchDarkly
+- Call OpenAI directly using your own client
+- Automatically track latency, token usage, and success/error via the tracker
+
+View your data in the LaunchDarkly dashboard under **Observability**.
+
+## Adapting for other providers
+
+To use a different provider, replace the OpenAI-specific parts:
+
+1. Swap the OpenAI client for your provider's client
+2. Use the appropriate tracker method (see table above), or use `trackMetricsOf` with a custom metrics extractor
+3. Map `aiConfig.messages` and `aiConfig.model` to your provider's API format
+
+See the [bedrock](../bedrock/) example for an AWS Bedrock adaptation.
diff --git a/packages/sdk/server-ai/examples/openai-observability/package.json b/packages/sdk/server-ai/examples/openai-observability/package.json
new file mode 100644
index 0000000000..acb4a22273
--- /dev/null
+++ b/packages/sdk/server-ai/examples/openai-observability/package.json
@@ -0,0 +1,23 @@
+{
+  "name": "openai-observability-example",
+  "version": "1.0.0",
+  "description": "LaunchDarkly AI SDK example: provider-specific observability with OpenAI",
+  "scripts": {
+    "build": "tsc",
+    "start": "yarn build && node ./dist/index.js"
+  },
+  "dependencies": {
+    "@launchdarkly/node-server-sdk": "workspace:^",
+    "@launchdarkly/observability-node": "^1.0.0",
+    "@launchdarkly/server-sdk-ai": "workspace:^",
+    "@launchdarkly/server-sdk-ai-openai": "workspace:^",
+    "@opentelemetry/instrumentation": "^0.57.0",
+    "@traceloop/instrumentation-openai": "^0.22.0",
+    "dotenv": "^16.0.0",
+    "openai": "^5.12.2"
+  },
+  "devDependencies": {
+    "@types/node": "^20.0.0",
+    "typescript": "^5.0.0"
+  }
+}
diff --git a/packages/sdk/server-ai/examples/openai-observability/src/index.ts b/packages/sdk/server-ai/examples/openai-observability/src/index.ts
new file mode 100644
index 0000000000..eb8d1f2470
--- /dev/null
+++ b/packages/sdk/server-ai/examples/openai-observability/src/index.ts
@@ -0,0 +1,99 @@
+/* eslint-disable no-console */
+import { registerInstrumentations } from '@opentelemetry/instrumentation';
+import { OpenAIInstrumentation } from '@traceloop/instrumentation-openai';
+import 'dotenv/config';
+
+import { init, type LDContext } from '@launchdarkly/node-server-sdk';
+import { Observability } from '@launchdarkly/observability-node';
+import { initAi } from '@launchdarkly/server-sdk-ai';
+
+const sdkKey = process.env.LAUNCHDARKLY_SDK_KEY;
+const aiConfigKey = process.env.LAUNCHDARKLY_AI_CONFIG_KEY || 'sample-ai-config';
+
+if (!sdkKey) {
+  console.error('*** Please set the LAUNCHDARKLY_SDK_KEY env first');
+  process.exit(1);
+}
+
+// ── 1. Initialize the LaunchDarkly client with the Observability plugin ──
+// The plugin automatically captures SDK operations, flag evaluations,
+// error monitoring, logging, and distributed tracing.
+const ldClient = init(sdkKey, {
+  plugins: [
+    new Observability({
+      serviceName: process.env.SERVICE_NAME || 'hello-js-openai-observability',
+      serviceVersion: process.env.SERVICE_VERSION || '1.0.0',
+    }),
+  ],
+});
+
+registerInstrumentations({
+  instrumentations: [new OpenAIInstrumentation()],
+});
+
+const context: LDContext = {
+  kind: 'user',
+  key: 'example-user-key',
+  name: 'Sandy',
+};
+
+async function main() {
+  try {
+    await ldClient.waitForInitialization({ timeout: 10 });
+    console.log('*** SDK successfully initialized');
+  } catch (error) {
+    console.error(`*** SDK failed to initialize: ${error}`);
+    process.exit(1);
+  }
+
+  const aiClient = initAi(ldClient);
+
+  // ── 2. Import provider and OpenAI after instrumentation so OpenLLMetry can patch the client ──
+  const { OpenAIProvider } = await import('@launchdarkly/server-sdk-ai-openai');
+  const { OpenAI } = await import('openai');
+  const openai = new OpenAI({
+    apiKey: process.env.OPENAI_API_KEY,
+  });
+
+  // ── 3. Get the AI Config (model, messages, parameters) from LaunchDarkly ──
+  // `completionConfig` returns the resolved configuration plus a `tracker`
+  // that you use to report metrics back to LaunchDarkly.
+  const aiConfig = await aiClient.completionConfig(
+    aiConfigKey,
+    context,
+    {
+      model: { name: 'gpt-4' },
+      enabled: false,
+    },
+    { example_type: 'provider_observability_demo' },
+  );
+
+  if (!aiConfig.enabled || !aiConfig.tracker) {
+    console.log('*** AI configuration is not enabled');
+    ldClient.close();
+    process.exit(0);
+  }
+
+  try {
+    // ── 4. Call OpenAI and track metrics with the provider's extractor ──
+    const completion = await aiConfig.tracker.trackMetricsOf(
+      OpenAIProvider.getAIMetricsFromResponse,
+      () =>
+        openai.chat.completions.create({
+          messages: aiConfig.messages || [],
+          model: aiConfig.model?.name || 'gpt-4',
+          temperature: (aiConfig.model?.parameters?.temperature as number) ?? 0.5,
+          max_tokens: (aiConfig.model?.parameters?.maxTokens as number) ?? 4096,
+        }),
+    );
+
+    console.log('AI Response:', completion.choices[0]?.message.content);
+    console.log('\nSuccess.');
+  } catch (err) {
+    console.error('Error:', err);
+  } finally {
+    ldClient.close();
+  }
+}
+
+main();
diff --git a/packages/sdk/server-ai/examples/openai-observability/tsconfig.json b/packages/sdk/server-ai/examples/openai-observability/tsconfig.json
new file mode 100644
index 0000000000..01907a899c
--- /dev/null
+++ b/packages/sdk/server-ai/examples/openai-observability/tsconfig.json
@@ -0,0 +1,18 @@
+{
+  "compilerOptions": {
+    "target": "ES2022",
+    "module": "CommonJS",
+    "moduleResolution": "node",
+    "esModuleInterop": true,
+    "allowSyntheticDefaultImports": true,
+    "strict": true,
+    "skipLibCheck": true,
+    "forceConsistentCasingInFileNames": true,
+    "outDir": "./dist",
+    "rootDir": "./src",
+    "declaration": true,
+    "sourceMap": true
+  },
+  "include": ["src/**/*"],
+  "exclude": ["node_modules", "dist"]
+}