Refactor-ai-inference

src/content/docs/en/pages/guides/ai-inference-starter-kit/index.mdx

@@ -0,0 +1,114 @@
+---
+title: Deploy AI Inference Starter Kit
+description: Deploy a complete AI Inference environment using the Starter Kit template on Azion.
+meta_tags: >-
+  ai inference, starter kit, template, deployment, artificial intelligence, edge computing
+namespace: docs_guides_ai_inference_starter_kit
+permalink: /documentation/products/guides/ai-inference-starter-kit/
+menu_namespace: AIInferenceMenu
+
+---
+
+import LinkButton from 'azion-webkit/linkbutton';
+
+The **AI Inference Starter Kit** provides a ready-to-use template for deploying AI capabilities on Azion's edge infrastructure. It includes pre-configured applications, functions, and model integrations.
+
+## Overview
+
+The Starter Kit includes:
+
+- An edge application configured for AI workloads
+- Pre-built functions for common AI tasks
+- Integration with available AI models
+- Sample code for quick implementation
+
+## Requirements
+
+Before you begin, ensure you have:
+
+- An Azion account
+- Access to Real-Time Manager
+
+## Deploy the Starter Kit
+
+1. Access the [Azion Console](https://console.azion.com/).
+2. On the **+ Create** page, search for **AI Inference Starter Kit**.
+3. Select the template.
+4. Configure your edge application:
+   - Enter a name for your application
+   - Select the models you want to use
+   - Configure additional settings as needed
+5. Click **Deploy**.
+
+<LinkButton
+    label="Deploy Starter Kit"
+    link="https://console.azion.com/create/azion/starter-kit-edge-ai"
+    icon="ai ai-azion"  
+    icon-pos="left"
+  />
+
+## Next steps
+
+After deployment, you can:
+
+- [Explore available models](/en/documentation/products/ai/ai-inference/models/)
+- [Build a simple agent](/en/documentation/products/guides/ai-inference-agent/)
+- [Integrate with SQL Database](/en/documentation/products/store/sql-database/) for vector search capabilities
+
+## Example usage
+
+After deploying, test your AI Inference with a simple request:
+
+```javascript
+const response = await Azion.AI.run("Qwen/Qwen3-30B-A3B-Instruct-2507-FP8", {
+  "stream": false,
+  "messages": [
+    {
+      "role": "user",
+      "content": "Explain edge computing in one sentence."
+    }
+  ]
+})
+
+console.log(response.choices[0].message.content)
+```
+
+Expected response:
+
+```json
+{
+  "id": "chatcmpl-abc123",
+  "choices": [{
+    "message": {
+      "role": "assistant",
+      "content": "Edge computing processes data closer to its source, reducing latency and bandwidth usage by bringing computation and storage resources near the devices or sensors that generate the data."
+    }
+  }]
+}
+```
+
+## Troubleshooting
+
+### Deployment fails
+
+If deployment fails, check:
+
+- Your account has sufficient permissions
+- All required fields are filled correctly
+- Error logs in the console for specific issues
+
+### Model not responding
+
+If the model isn't responding:
+
+- Verify the function is correctly configured
+- Confirm the model name is correct in `Azion.AI.run()`
+- Check the rate limits in the [Limits section](/en/documentation/products/ai/ai-inference/#limits)
+
+### High latency
+
+If you experience high latency:
+
+- Consider using streaming responses (`"stream": true`)
+- Check if your edge application is deployed in the optimal region
+- Review your function's timeout settings