<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Sarfaraz Merchant on Medium]]></title>
        <description><![CDATA[Stories by Sarfaraz Merchant on Medium]]></description>
        <link>https://medium.com/@sarfarazMerchant?source=rss-5258823104b4------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*oFnRknrXPUGeg-xZiZktXg.jpeg</url>
            <title>Stories by Sarfaraz Merchant on Medium</title>
            <link>https://medium.com/@sarfarazMerchant?source=rss-5258823104b4------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 01:02:33 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@sarfarazMerchant/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Beyond Chat: Processing Images, PDFs, and Documents with the OpenAI Adapter in Oracle Integration…]]></title>
            <link>https://pub.towardsai.net/beyond-chat-processing-images-pdfs-and-documents-with-the-openai-adapter-in-oracle-integration-86253d02f84b?source=rss-5258823104b4------2</link>
            <guid isPermaLink="false">https://medium.com/p/86253d02f84b</guid>
            <category><![CDATA[oracle-integration-cloud]]></category>
            <category><![CDATA[openai]]></category>
            <category><![CDATA[image-processing]]></category>
            <category><![CDATA[oracle]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Sarfaraz Merchant]]></dc:creator>
            <pubDate>Mon, 08 Jun 2026 13:28:36 GMT</pubDate>
            <atom:updated>2026-06-18T06:08:46.343Z</atom:updated>
            <content:encoded><![CDATA[<h3>Beyond Chat: Processing Images, PDFs, and Documents with the OpenAI Adapter in Oracle Integration Cloud</h3><h4><strong><em>Exploring File Uploads, Image Processing, and Document Extraction using the Native OpenAI Adapter in OIC.</em></strong></h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*i8fm4YavzSzD_aYmwBL4Wg.png" /></figure><p>When Oracle introduced the OpenAI Adapter in Oracle Integration Cloud (OIC), most of the examples I came across focused on text generation and chatbot-style interactions.</p><p>Naturally, I assumed that if I wanted to process files such as images, PDFs, or Word documents, I would probably need to build custom REST integrations and call the OpenAI APIs directly.</p><p>While exploring the adapter, I noticed a few operations related to file management. That got me curious.</p><p>Can the OpenAI Adapter <strong>upload files</strong>?</p><p>Can those files be <strong>processed</strong> later using a File ID?</p><p>Can we extract structured data from <strong>invoices</strong>, <strong>images</strong>, or <strong>PDF </strong>documents <strong>without </strong>building custom REST calls?</p><p>To answer these questions, I decided to build a small proof of concept.</p><p>The results were quite interesting.</p><p>Using only the native OpenAI Adapter, I was able to:</p><ul><li>Upload files to OpenAI</li><li>Retrieve and reuse File IDs</li><li>Process images and PDF documents</li><li>Extract structured JSON responses</li><li>Analyze multilingual content, including Arabic and Persian documents</li></ul><p>In this article, I’ll walk through the approach, the architecture, and a few practical use cases that I tested while building the POC.</p><p>By the end, you’ll have a reusable pattern that can be applied to invoice extraction, document processing, content analysis, and many other AI-powered integration scenarios.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/656/1*AmnPbjUcPwRDhJIX7R6wbw.png" /><figcaption><em>OpenAI Adapter provides native operations for uploading and processing files.</em></figcaption></figure><h3>Understanding the Architecture</h3><p>While building the POC, I decided to separate the solution into two integrations.</p><p>The first integration is responsible for uploading a file to OpenAI and obtaining a File ID.</p><p>The second integration uses that File ID together with a prompt to process the document and return a structured response.</p><p>I found this approach cleaner because the uploaded file becomes a reusable asset. The same file can be analyzed multiple times using different prompts without uploading it again.</p><h3>Integration 1 — Upload File</h3><p>The first integration accepts a file and uploads it using the OpenAI Adapter’s <strong>Upload File</strong> operation.</p><p>Supported examples include:</p><ul><li>Images (JPG, PNG)</li><li>PDF documents</li><li>Microsoft Word documents</li></ul><p>The adapter returns a unique File ID similar to:</p><pre>file-xxxxxxxxxxxxxxxx</pre><p>This File ID can then be stored, logged, or passed to another integration for processing.</p><h3>Integration 2 — Process File</h3><p>The second integration receives:</p><ul><li>A File ID</li><li>Processing instructions</li></ul><p>It then invokes the OpenAI Adapter using the Responses operation.</p><p>Depending on the type of document and the prompt provided, OpenAI can:</p><ul><li>Extract invoice data</li><li>Analyze documents</li><li>Summarize content</li><li>Return structured JSON</li><li>Process multilingual text</li></ul><p>The response can be consumed directly within OIC and mapped to downstream systems.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uUiOlAR_PhDgtKSL64sFMw.png" /><figcaption>Solution Architecture.</figcaption></figure><h3>High-Level Flow</h3><p>The overall process looks like this:</p><pre>Document<br>(Image/PDF/DOCX)<br>        |<br>        v<br>+------------------+<br>| Upload File      |<br>| OpenAI Adapter   |<br>+------------------+<br>        |<br>        v<br>     File ID<br>        |<br>        v<br>+------------------+<br>| Responses API    |<br>| OpenAI Adapter   |<br>+------------------+<br>        |<br>        v<br> Structured Output<br>      (JSON)</pre><p>One thing I particularly liked about this pattern is that it relies entirely on the native OpenAI Adapter. There are no custom REST calls, no manual API payload construction, and no need to manage Base64-encoded content.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*gND494Sp2JCNGw5Dc775Ww.png" /><figcaption>Upload FIle to OpenAI Account.</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*cWttrE5HtVI-llOw0XSEGQ.png" /><figcaption>Process OpenAI Files.</figcaption></figure><h3>Building Integration 1 — Uploading Files to OpenAI</h3><p>The first integration is responsible for uploading a document to OpenAI and obtaining a File ID that can be reused later.</p><p>In my case, I tested the following file types:</p><ul><li>JPG images</li><li>PNG images</li><li>PDF documents</li><li>Microsoft Word documents</li></ul><h4>Step 1 — Create the OpenAI Connection</h4><p>Start by creating a connection using the OpenAI Adapter.</p><p>Provide:</p><ul><li>OpenAI API Key</li><li>Connection Name</li><li>Security Configuration</li></ul><p>Once the connection is successfully tested, it can be used within your integration.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-ZLdWYtfliQTkdynsmyJgA.png" /><figcaption>OpenAI Connection.</figcaption></figure><h3>Step 2 — Create the Upload File Integration</h3><p>Create an App-Driven Orchestration.</p><p>2.1 <strong>Add Trigger Rest Adapter to Accept File as input with</strong>:</p><p>- Select the multipart attachment processing options</p><ul><li>Request is multipart with payload</li><li>Multipart request is of type multipart/form-data with HTML form payload</li></ul><p>2.2 <strong>Add the OpenAI Adapter and select the following operation</strong>:</p><pre>Upload File</pre><p>This operation allows OIC to upload a file directly to OpenAI without invoking the REST API manually.</p><h3>Step 3 — Configure the Request Mapping</h3><p>The Upload File operation expects two important pieces of information:</p><h4>Purpose</h4><p>The purpose determines how the uploaded file will be used.</p><p>For image processing:</p><pre>vision</pre><p>For document processing (PDF, DOCX, TXT, etc.):</p><pre>user_data</pre><h4>File Content</h4><p>Map the incoming attachment or file reference to the adapter’s File element.</p><p>Example mapping:</p><pre>RequestWrapper<br> ├── Purpose<br> └── File<br>      └── streamReference</pre><p>The adapter handles the upload process and stores the file within OpenAI.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*N7z1X_bun5DUeeZ_I6LiTA.png" /><figcaption>Upload File Mapper.</figcaption></figure><h3>Step 4 — Execute the Integration</h3><p>After activation, invoke the integration with a sample document.</p><p>If the upload succeeds, the response contains a File ID.</p><p>Example:</p><pre>file-3UoRxBRyqV7pJRgPC4WSgi</pre><p>This File ID becomes the bridge between the upload step and the document processing step.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Bk9S2ZyY1Qm3zSilrpSQ7Q.png" /><figcaption>Uploaded File to OpenAI Accuont.</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*xHxziHxE6D9_nnB_ZzLeOw.png" /><figcaption>Storage FIles in OpenAI Account.</figcaption></figure><h3>Lesson Learned: File Names Matter</h3><p>While testing the Upload File operation, I ran into an issue that took a little time to troubleshoot.</p><p>Some uploads failed when the file name contained spaces or special characters.</p><p>For example:</p><pre>WhatsApp Image 2026-06-05 at 7.25.23 PM.jpeg</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*e0TkOGUr2lVEqhyP6d_LDw.png" /><figcaption>Error if the File Name contains space.</figcaption></figure><p>After renaming the file to a simpler format without spaces or special characters, the upload completed successfully.</p><p>Example:</p><pre>invoice_20260605.jpeg</pre><p>If you encounter unexpected upload failures, one of the first things to check is the file name.</p><p>Although this may vary depending on the adapter version and environment, using clean file names is a good practice and helped avoid issues during testing.</p><h3>Why the File ID Matters</h3><p>Initially, I assumed I would need to send Base64 content to OpenAI every time I wanted to analyze a document.</p><p>The File ID approach is much cleaner.</p><p>Once the document is uploaded:</p><ul><li>The file can be reused</li><li>Multiple prompts can be executed against the same file</li><li>There is no need to repeatedly transfer the document</li></ul><p>This makes the pattern both efficient and easy to maintain.</p><p>At this stage, we have successfully uploaded a document and obtained a File ID.</p><p>The next step is where things become interesting: using that File ID to extract meaningful information from the document.</p><blockquote>“This was the point where I realized the adapter was capable of much more than chat-based interactions. The File ID effectively turns the uploaded document into a reusable asset that can be processed multiple times.”</blockquote><h3>Building Integration 2 — Processing Files Using the Responses API</h3><p>Now that the document has been uploaded and we have a File ID, the next step is to ask OpenAI to process the file.</p><p>This is where the Responses operation comes into play.</p><p>The Responses API allows us to provide:</p><ul><li>Instructions (prompt)</li><li>File reference (File ID)</li><li>Expected output format</li></ul><p>and receive a structured response.</p><h3>Step 1 — Create the Processing Integration</h3><p>Create a second integration.</p><p>This integration accepts:</p><p><strong>FileId</strong> — Contains the OpenAI File ID returned by the Upload File integration.</p><p><strong>FileType</strong> — Determines the type of file being processed. For images, I passed input_image, while for PDFs and Word documents, I passed input_file.</p><p><strong>instruction</strong> — Contains the prompt or instructions that should be sent to the model.</p><p>Add the OpenAI Adapter and select:</p><pre>Responses</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/695/1*UMjWJJ5BCDVf0YKgGbg2IA.png" /><figcaption>OpenAI Responses.</figcaption></figure><h3>Step 2 — Build the Request</h3><p>The request consists of two content elements:</p><h4>Content 1 — Instructions</h4><p>The first content item contains the instructions we want the model to follow.</p><p>Example:</p><pre>Extract invoice information from the attached document.<br>Return valid JSON only.<br>Include:<br>- Invoice Number<br>- Date<br>- Items<br>- Tax<br>- Total Amount</pre><p>In the mapper:</p><pre>Type = input_text<br>Text = &lt;your prompt&gt;</pre><h4>Content 2 — File Reference</h4><p>The second content item references the uploaded file.</p><p>For image files:</p><pre>Type = input_image<br>File Id = file-xxxxxxxx</pre><p>For PDF and Word documents:</p><pre>Type = input_file<br>File Id = file-xxxxxxxx</pre><p>The File ID comes from the Upload File integration we created earlier.</p><p>At runtime, OpenAI retrieves the uploaded file and processes it together with the prompt.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*sG0dG9Sg08mfUcuDe9TnGw.png" /><figcaption>Responses Mapping.</figcaption></figure><p>Inside the Responses request, I configured the following mappings:</p><ul><li><strong>Model</strong> → gpt-4.1</li><li><strong>Role</strong> → user</li><li><strong>Content 1 Type</strong> → input_text</li><li><strong>Content 1 Text</strong> → Query Parameter instruction</li><li><strong>Content 2 Type</strong> → Query Parameter FileType</li><li><strong>Content 2 File ID</strong> → Query Parameter FileId</li></ul><p>This design allowed me to use the same integration for multiple scenarios simply by changing the values passed at runtime.</p><p>For example, when processing a restaurant invoice image, I passed:</p><ul><li>FileType = input_image</li><li>FileId = OpenAI File ID</li><li>instruction = “Extract invoice details and return valid JSON”</li></ul><p>For a PDF document, I only changed the FileType and instruction:</p><ul><li>FileType = input_file</li><li>FileId = OpenAI File ID</li><li>instruction = “Extract page-wise content and return valid JSON”</li></ul><p>The integration logic remained exactly the same.</p><h3>Step 3 — Configure the Model</h3><p>For my testing, I used GPT-4.1.</p><p>Recommended settings:</p><pre>Temperature = 0</pre><p>When extracting structured information, deterministic responses are generally easier to consume downstream.</p><h3>Step 4 — Execute the Integration</h3><p>After activation, invoke the integration using:</p><ul><li>File ID</li><li>Processing prompt</li><li>File Type</li></ul><p>The model analyzes the file and returns a response.</p><p>A typical response might look like:</p><pre>{<br>  &quot;Result&quot; : &quot;```json\n{\n  \&quot;Invoice Number\&quot;: \&quot;123982857\&quot;,\n  \&quot;Date\&quot;: \&quot;23/04/2022\&quot;,\n  \&quot;Items\&quot;: [\n    {\&quot;Item Name\&quot;: \&quot;شاورما دجاج عربية عصملية\&quot;, \&quot;Qty\&quot;: 1, \&quot;Total\&quot;: 8.000},\n    {\&quot;Item Name\&quot;: \&quot;حمص\&quot;, \&quot;Qty\&quot;: 2, \&quot;Total\&quot;: 4.000},\n    {\&quot;Item Name\&quot;: \&quot;فتوش\&quot;, \&quot;Qty\&quot;: 1, \&quot;Total\&quot;: 3.000},\n    {\&quot;Item Name\&quot;: \&quot;تبولة\&quot;, \&quot;Qty\&quot;: 1, \&quot;Total\&quot;: 2.500},\n    {\&quot;Item Name\&quot;: \&quot;بطاطا مقلية\&quot;, \&quot;Qty\&quot;: 1, \&quot;Total\&quot;: 2.000},\n    {\&quot;Item Name\&quot;: \&quot;خبز صاج\&quot;, \&quot;Qty\&quot;: 2, \&quot;Total\&quot;: 1.000},\n    {\&quot;Item Name\&quot;: \&quot;شاورما لحم عربية\&quot;, \&quot;Qty\&quot;: 1, \&quot;Total\&quot;: 8.000},\n    {\&quot;Item Name\&quot;: \&quot;كولا زجاج\&quot;, \&quot;Qty\&quot;: 2, \&quot;Total\&quot;: 4.000},\n    {\&quot;Item Name\&quot;: \&quot;لبن\&quot;, \&quot;Qty\&quot;: 2, \&quot;Total\&quot;: 4.000},\n    {\&quot;Item Name\&quot;: \&quot;مية 250 مل\&quot;, \&quot;Qty\&quot;: 2, \&quot;Total\&quot;: 2.000}\n  ],\n  \&quot;Tax\&quot;: 2.420,\n  \&quot;Total Amount\&quot;: 46.570\n}\n```&quot;<br>}</pre><p>What impressed me most was that the model not only extracted the invoice totals, but also correctly identified <strong>Arabic </strong>item names, quantities, and line totals, and returned them in a clean JSON structure.</p><p>Within OIC, this response can be parsed and then can be mapped directly to downstream applications or business processes.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6bcwUIKAprWoh-rvv2wfkw.png" /><figcaption>Processed Image Data.</figcaption></figure><h3>Use Case 1 — Restaurant Invoice Extraction</h3><p>To validate the solution, I started with a restaurant invoice written in Arabic.</p><p>The objective was simple:</p><ul><li>Upload the invoice image</li><li>Process it using the OpenAI Adapter</li><li>Extract business data in a structured format</li></ul><p>Rather than returning a textual description of the invoice, the model was instructed to generate JSON containing:</p><ul><li>Invoice Number</li><li>Invoice Date</li><li>Item Details</li><li>Tax Amount</li><li>Total Amount</li></ul><p>The result was surprisingly accurate.</p><p>The model successfully identified:</p><ul><li>Arabic item names</li><li>Item quantities</li><li>Line totals</li><li>Tax information</li><li>Grand total</li></ul><p>This demonstrates how the OpenAI Adapter can be used as a lightweight document extraction service directly within Oracle Integration Cloud.</p><p>Potential use cases include:</p><ul><li>Supplier invoice processing</li><li>Expense receipts</li><li>Restaurant bills</li><li>Retail receipts</li><li>Purchase orders</li></ul><p>For organizations already using OIC, this pattern can significantly reduce manual data entry and simplify document-driven workflows.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*8tRdngr6IEyuTjQ-MF7UMg.gif" /></figure><h3>Use Case 2 — Processing Persian and Arabic Documents</h3><p>After validating the invoice extraction scenario, I wanted to see how the adapter would perform with larger documents.</p><p>For this test, I used Persian and Arabic PDF documents containing multiple pages of text.</p><p>The goal was different from the invoice use case.</p><p>Instead of extracting business fields, I wanted to:</p><ul><li>Preserve the original language</li><li>Maintain page boundaries</li><li>Separate paragraphs</li><li>Return a structured response suitable for further processing</li></ul><p>The prompt instructed the model to return page-wise JSON.</p><p>A simplified response structure looked like this:</p><pre>{<br>  &quot;pages&quot;: [<br>    {<br>      &quot;page_number&quot;: 1,<br>      &quot;paragraphs&quot;: [<br>        &quot;...&quot;,<br>        &quot;...&quot;<br>      ]<br>    }<br>  ]<br>}</pre><p>The results were encouraging.</p><p>The model was able to:</p><ul><li>Read multilingual content</li><li>Separate pages</li><li>Preserve paragraph structure</li><li>Return machine-readable JSON</li></ul><p>This opens up interesting possibilities for:</p><ul><li>Digitizing legacy documents</li><li>Knowledge extraction</li><li>Content indexing</li><li>Search and retrieval solutions</li><li>Document migration projects</li></ul><p>One important observation is that document understanding and transcription accuracy are two different goals.</p><p>For scenarios where verbatim OCR accuracy is critical, Oracle Document Understanding may still be a valuable companion service. However, for document analysis and content structuring, the OpenAI Adapter performed remarkably well.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*fNNT7A77tE9HBBWkVuaqug.gif" /></figure><h3>Lessons Learned</h3><ul><li><strong>Use clean file names</strong> — avoid spaces and special characters when uploading files.</li><li><strong>Upload once, reuse many times</strong> — leverage the returned File ID instead of sending file content repeatedly.</li><li><strong>Parameterize your integration</strong> — File ID, File Type, and Instructions make the solution highly reusable.</li><li><strong>Prompt engineering matters</strong> — clear instructions significantly improve extraction quality and consistency.</li><li><strong>Expect an additional parsing step</strong> — JSON responses are returned inside the Result field.</li><li><strong>One integration can support multiple document types</strong> — images, PDFs, and Word documents can be processed using the same pattern.</li><li><strong>The OpenAI Adapter is more than chat</strong> — it can extract, understand, and structure information from documents and images.</li></ul><h3>Source Code</h3><p>The complete Oracle Integration Cloud package (.car), sample prompts, and supporting assets used in this article are available on GitHub.</p><p>GitHub Repository:</p><p><a href="https://github.com/sarfarazmerchant/oic-openai-adapter-file-processing.git">https://github.com/sarfarazmerchant/oic-openai-adapter-file-processing.git</a></p><p>Feel free to download the package, import it into your OIC environment, and adapt it to your own document-processing use cases.</p><h3>Final Thoughts</h3><p>This proof of concept demonstrated that the OpenAI Adapter in Oracle Integration Cloud can do much more than chat interactions. By combining the <strong>Upload File</strong> and <strong>Responses</strong> operations, I was able to process images, PDFs, and Word documents, and transform them into structured JSON ready for downstream integrations.</p><p>The pattern is simple:</p><pre>Upload File<br>    ↓<br>Get File ID<br>    ↓<br>Process with Responses API<br>    ↓<br>Receive Structured Data</pre><p>For me, this was a practical introduction to AI-powered document processing in OIC, and I hope it helps other integration developers explore similar use cases.</p><blockquote><strong><em>Have you explored the OpenAI Adapter in OIC? I’d love to hear about your use cases and experiences.</em></strong></blockquote><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=86253d02f84b" width="1" height="1" alt=""><hr><p><a href="https://pub.towardsai.net/beyond-chat-processing-images-pdfs-and-documents-with-the-openai-adapter-in-oracle-integration-86253d02f84b">Beyond Chat: Processing Images, PDFs, and Documents with the OpenAI Adapter in Oracle Integration…</a> was originally published in <a href="https://pub.towardsai.net/">Towards AI</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Oracle APEX 26.1: APEXLang, AI Development & the Future of Oracle Low-Code]]></title>
            <link>https://medium.com/@sarfarazMerchant/oracle-apex-26-1-apexlang-ai-development-the-future-of-oracle-low-code-0f0086ec2e04?source=rss-5258823104b4------2</link>
            <guid isPermaLink="false">https://medium.com/p/0f0086ec2e04</guid>
            <category><![CDATA[generative-ai-tools]]></category>
            <category><![CDATA[oracle-apex]]></category>
            <category><![CDATA[apexlang]]></category>
            <category><![CDATA[ace-oracle]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Sarfaraz Merchant]]></dc:creator>
            <pubDate>Fri, 22 May 2026 06:34:56 GMT</pubDate>
            <atom:updated>2026-05-22T06:42:09.639Z</atom:updated>
            <content:encoded><![CDATA[<h4>A deep dive into the biggest announcements from Oracle APEX 26.1 — including APEXLang, AI-assisted development, VS Code workflows, Git integration, and what it means for Oracle developers.</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*hQQd1MTlH9OdukOt2MFfpA.jpeg" /></figure><h3>Oracle APEX 26.1 feels different.</h3><p>This release is not just about adding a few new features or UI improvements. It represents a major shift in how Oracle envisions the future of application development — especially around:</p><ul><li>AI-assisted development</li><li>Git-native workflows</li><li>VS Code integration</li><li>Team-based DevOps</li><li>Source-controlled applications</li></ul><p>During the recent Oracle APEX 26.1 Session Part 2 Q&amp;A, the Oracle APEX product team shared several important insights about the future direction of the platform.</p><p>And one topic clearly dominated the discussion:</p><h3>APEXLang</h3><h3>What Is APEXLang?</h3><p>APEXLang is Oracle’s new human-readable export/import format for Oracle APEX applications.</p><p>Instead of relying only on large SQL export files, Oracle now provides a format specifically designed for:</p><ul><li>Readability</li><li>Version control</li><li>AI-assisted development</li><li>Easier Git diffs</li><li>VS Code workflows</li></ul><p>According to the Oracle APEX team, APEXLang was intentionally designed to be:</p><ul><li>Easier to read than JSON</li><li>Less verbose than YAML</li><li>More developer-friendly for AI tools</li></ul><p>This is a huge step toward treating APEX applications like modern source-controlled projects.</p><h3>Why This Matters</h3><p>Traditionally, Oracle APEX applications lived mostly inside the database.</p><p>That worked well for rapid development, but it created challenges for:</p><ul><li>Git workflows</li><li>Merge handling</li><li>Multi-developer collaboration</li><li>CI/CD automation</li><li>AI-assisted coding</li></ul><p>APEXLang changes that.</p><p>Now developers can:</p><ul><li>Export applications into readable files</li><li>Open them in VS Code</li><li>Use AI coding assistants</li><li>Store them in Git</li><li>Automate deployments through CI/CD pipelines</li></ul><p>In many ways, Oracle APEX is moving closer to modern software engineering practices while still keeping its low-code strengths.</p><h3>AI Is Becoming a First-Class Citizen in APEX</h3><p>One of the most exciting discussions during the session was Oracle’s growing investment in AI-assisted development.</p><p>The Oracle APEX team confirmed that tools like:</p><ul><li>Claude</li><li>Codex</li><li>OpenAI-compatible models</li></ul><p>can already work with APEXLang and APEX Skills to generate application artifacts.</p><p>This means developers can potentially generate:</p><ul><li>Database schemas</li><li>SQL DDL</li><li>APEX components</li><li>Business logic</li><li>Templates</li><li>Entire application structures using AI coding agents.</li></ul><p>That’s a major shift for Oracle developers.</p><h3>The Recommended AI Workflow</h3><p>Oracle demonstrated a workflow that looks something like this:</p><pre>Export App → Open in VS Code → Use AI Agent → Validate with SQLcl → Import Back into APEX</pre><p>The process may still feel slightly manual today, but Oracle clearly indicated that tighter integrations are coming in future releases.</p><p>The direction is obvious:</p><p>Oracle wants APEX to become AI-native.</p><h3>VS Code Integration Is a Big Deal</h3><p>Oracle also emphasized that the future APEX developer experience will heavily involve:</p><ul><li>VS Code</li><li>SQLcl</li><li>Git</li><li>AI coding assistants</li></ul><p>To support this, Oracle now provides:</p><ul><li>Oracle SQL Developer Extension for VS Code</li><li>APEXLang support</li><li>Validation tooling</li><li>Auto-completion</li><li>Error highlighting</li></ul><p>This creates a much more modern development workflow compared to traditional browser-only development.</p><h3>CI/CD &amp; Git Workflows Are Now Officially Encouraged</h3><p>Another major theme was DevOps.</p><p>Oracle openly discussed workflows involving:</p><pre>VS Code → Git → CI/CD → APEX Environment</pre><p>This is important because many enterprises previously struggled with:</p><ul><li>Source control</li><li>Team collaboration</li><li>Environment synchronization</li><li>Automated deployments</li></ul><p>Oracle now appears fully committed to supporting enterprise-grade DevOps workflows.</p><h3>Current Limitations of APEXLang</h3><p>While APEXLang is exciting, Oracle was also transparent about current limitations.</p><p>At the moment:</p><ul><li>Only full application imports are supported</li><li>Partial imports are not available</li><li>Single-page imports are not supported</li><li>Full imports replace the target application</li><li>Some older component types are unsupported</li></ul><p>Oracle confirmed that incremental imports and better team workflows are already under investigation for future releases.</p><h3>Working Copies &amp; Team Collaboration</h3><p>One of the most common concerns raised during the session was:</p><p>“How do multiple developers work on the same application?”</p><p>Oracle acknowledged that this is still evolving.</p><p>The team confirmed:</p><ul><li>Working copies will play a major role</li><li>Better integration is coming</li><li>Future patches will improve merge handling</li><li>Upcoming Office Hours sessions will focus entirely on team workflows</li></ul><p>This is clearly an area Oracle is actively investing in.</p><h3>AI Features Beyond APEXLang</h3><p>APEX 26.1 also expands AI capabilities across the platform itself.</p><p>New and evolving AI features include:</p><ul><li>AI-assisted page creation</li><li>AI-generated applications</li><li>AI Search for Interactive Reports</li><li>OCI Generative AI integration</li><li>OpenAI-compatible APIs</li><li>APEX_AI.GENERATE enhancements</li></ul><p>Oracle also hinted that future releases will significantly expand these capabilities.</p><h3>Translation Improvements</h3><p>Another underrated improvement in APEX 26.1 is the simplified translation model.</p><p>Oracle introduced a text-message-based translation approach that reduces dependency on older “shadow application” translation methods.</p><p>This could make multilingual applications significantly easier to maintain.</p><h3>My Thoughts on the Direction of Oracle APEX</h3><p>After watching the session, one thing became very clear:</p><p>Oracle APEX is no longer positioning itself as just a low-code platform.</p><p>It’s becoming:</p><ul><li>AI-aware</li><li>Git-native</li><li>VS Code-friendly</li><li>CI/CD-ready</li><li>Enterprise DevOps-compatible</li></ul><p>And APEXLang is the foundation for that transition.</p><p>There are still rough edges today — especially around partial imports and collaborative workflows — but the roadmap Oracle shared looks extremely promising.</p><p>For developers already invested in Oracle technologies, APEX 26.1 may become one of the most important releases in recent years.</p><h3>Useful Resources</h3><ul><li>Oracle APEX: <a href="https://www.oracle.com/apex/">https://www.oracle.com/apex/</a></li><li>APEX Updates: <a href="https://apex.oracle.com/go/adb-updates">https://apex.oracle.com/go/adb-updates</a></li><li>Oracle APEX Forums: <a href="https://apex.oracle.com/forum">https://apex.oracle.com/forum</a></li><li>SQL Developer for VS Code: <a href="https://www.oracle.com/database/sqldeveloper/vscode/">https://www.oracle.com/database/sqldeveloper/vscode/</a></li><li>APEX Skills GitHub: <a href="https://github.com/oracle/skills">https://github.com/oracle/skills</a></li><li>AI App Generator: <a href="https://www.oracle.com/apex/ai-application-generator/">https://www.oracle.com/apex/ai-application-generator/</a></li></ul><h3>Final Thoughts</h3><p>Oracle APEX 26.1 feels like the beginning of a major transformation.</p><p>The combination of:</p><ul><li>APEXLang</li><li>AI-assisted development</li><li>VS Code integration</li><li>Git workflows</li><li>CI/CD support</li></ul><p>suggests Oracle is aggressively modernizing the APEX developer experience.</p><p>The next 12–18 months for Oracle APEX are going to be very interesting.</p><h4>If you’re already experimenting with APEX 26.1 or APEXLang, I’d love to hear your thoughts and experiences.</h4><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=0f0086ec2e04" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Oracle APEX 26.1: The Ultimate Resource Hub & Why ‘Vibe Coding’ Just Went Mainstream]]></title>
            <link>https://medium.com/@sarfarazMerchant/oracle-apex-26-1-the-ultimate-resource-hub-why-vibe-coding-just-went-mainstream-72573aa9f4e8?source=rss-5258823104b4------2</link>
            <guid isPermaLink="false">https://medium.com/p/72573aa9f4e8</guid>
            <category><![CDATA[apexlang]]></category>
            <category><![CDATA[vibe-coding]]></category>
            <category><![CDATA[oracle-apex]]></category>
            <category><![CDATA[oracle]]></category>
            <category><![CDATA[ace-oracle]]></category>
            <dc:creator><![CDATA[Sarfaraz Merchant]]></dc:creator>
            <pubDate>Fri, 15 May 2026 13:28:25 GMT</pubDate>
            <atom:updated>2026-05-15T13:49:07.595Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*PUNHheofwvdMAlbXDFKXxw.jpeg" /></figure><p>The buzz around the May 14th release of <strong>Oracle APEX 26.1</strong> is deafening. Everyone is talking about the new AI features, but as developers, we know that information is currently scattered across blogs, documentation, and videos.</p><p>I wanted to do something different. Instead of just adding to the noise, I’ve compiled <strong>every critical resource and update</strong> into this single jumping-off point. Whether you are looking for the “how-to” videos or the deep-dive documentation, you’ll find it below.</p><h3>🏗️ The Paradigm Shift: APEX enters the “Vibe Coding” Era</h3><p>We’ve moved past simple low-code. With this release, Oracle is leaning into what the industry is calling <strong>“Vibe Coding”</strong> — the ability to describe your intent and have the platform handle the structural heavy lifting.</p><h3>The AI Trifecta:</h3><ul><li><strong>APEXlang:</strong> This is the secret sauce. It’s a human-readable metadata spec that acts as the bridge between your natural language “vibe” and a functional app structure. (Explore here:- <a href="https://www.oracle.com/apex/ai-application-generator/">ApexLang</a>)</li><li><strong>AI Agents:</strong> These aren’t just chatbots. They are action-oriented entities that can execute PL/SQL and JavaScript within the boundaries you set.</li><li><strong>AI Interactive Reports:</strong> Transforming user intent (e.g., <em>“Show me our highest-growth regions”</em>) into declarative report settings. No SQL required for the end-user, but complete control for the dev.</li></ul><h3>📚 The 26.1 Knowledge Vault (All Resources in One Place)</h3><p>If you are looking to master this release, start here. <strong>Bookmark this list.</strong></p><h3>🔗 Official Documentation &amp; Downloads</h3><ul><li><a href="https://www.oracle.com/apex/whats-new/"><strong>Official Release Home Page</strong></a><strong>:</strong> The high-level overview of every new feature.</li><li><a href="https://docs.oracle.com/en/database/oracle/apex/26.1/htmrn/index.html"><strong>Technical Release Notes</strong></a><strong>:</strong> The “under the hood” details, including bug fixes and compatibility.</li></ul><h3>🎥 Video Training</h3><ul><li><a href="https://www.youtube.com/watch?v=VIENX8sIO0c&amp;list=PLUo-NIMouZ_uRnUYrDLtY_nZ31a2wxOQs"><strong>APEX 26.1 YouTube Playlist</strong></a><strong>:</strong> Visual walkthroughs of the new AI Agents and Data Reporter.</li></ul><h3>✍️ Expert Deep-Dives (The Best Blog Posts)</h3><ul><li><strong>AI &amp; Agents:</strong></li><li><a href="https://blogs.oracle.com/apex/ai-agents-in-oracle-apex"><strong><em>Move from Insights to Action with AI Agents</em></strong></a></li><li><a href="https://blogs.oracle.com/apex/build-a-crm-ai-agent-with-oracle-apex"><strong><em>Step-by-Step: Build a CRM AI Agent</em></strong></a></li><li><strong>Reporting &amp; Self-Service:</strong></li><li><a href="https://blogs.oracle.com/apex/oracle-apex-data-reporter"><strong><em>Putting Reporting Back in the Hands of Business Users</em></strong></a></li><li><a href="https://blogs.oracle.com/apex/introducing-apex-ai-interactive-reports"><strong><em>A Deep Dive into AI Interactive Reports</em></strong></a></li><li><a href="https://blogs.oracle.com/apex/oracle-apex-data-reporter"><strong><em>Meet the Oracle APEX Data Reporter</em></strong></a></li></ul><p><strong>Oracle APEX Newsletter:</strong></p><ul><li><a href="https://oracleapex.com/ords/r/apex_pm/newsletter-app/newsletter?edition=26.1"><strong><em>Special Edition: Oracle APEX 26.1</em></strong></a></li></ul><h3>📅 Must-Attend: Release Office Hours</h3><p>To truly master these updates, you must attend the official <strong>Release Office Hours</strong>. These sessions are the best way to see live demos and ask the product team direct questions. <strong>Make sure to register at the links below:</strong></p><ul><li><strong>May 21, 2026:</strong> <a href="https://asktom.oracle.com/ords/r/tech/catalog/session-landing-page?p2_event_id=108686086326684389626004407045681987216&amp;p2_source_log=QR">Part 2: What’s new in Oracle APEX 26.1</a></li><li><strong>May 28, 2026:</strong> <a href="https://asktom.oracle.com/ords/r/tech/catalog/session-landing-page?p2_event_id=108689880475038199439393107080286760229&amp;p2_source_log=QR">Part 3: What’s new in Oracle APEX 26.1</a></li><li><strong>June 04, 2026:</strong> <a href="https://asktom.oracle.com/ords/r/tech/catalog/session-landing-page?p2_event_id=108689977357298124411424273388932852675&amp;p2_source_log=QR">Part 4: What’s new in Oracle APEX 26.1</a></li><li><strong>June 11, 2026:</strong> <a href="https://asktom.oracle.com/ords/r/tech/catalog/session-landing-page?p2_event_id=108690669561065422741464641540591352473&amp;p2_source_log=QR">Part 5: What’s new in Oracle APEX 26.1</a></li><li><strong>June 25, 2026:</strong> <a href="https://asktom.oracle.com/ords/r/tech/catalog/session-landing-page?p2_event_id=108690025403004694895583407695603592040&amp;p2_source_log=QR">Part 6: What’s new in Oracle APEX 26.1</a></li></ul><p><strong>What is the first feature you’re going to implement in your current production apps?</strong> Let’s share our findings in the comments so we can all learn this new era of APEX together.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=72573aa9f4e8" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Email Trick: How I Automated APEX App Security Without a User Table]]></title>
            <link>https://medium.com/@sarfarazMerchant/the-email-trick-how-i-automated-apex-app-security-without-a-user-table-c4c4862cdf20?source=rss-5258823104b4------2</link>
            <guid isPermaLink="false">https://medium.com/p/c4c4862cdf20</guid>
            <category><![CDATA[oracle]]></category>
            <category><![CDATA[authorization]]></category>
            <category><![CDATA[ace-oracle]]></category>
            <category><![CDATA[oracle-apex]]></category>
            <dc:creator><![CDATA[Sarfaraz Merchant]]></dc:creator>
            <pubDate>Sat, 09 May 2026 08:13:36 GMT</pubDate>
            <atom:updated>2026-05-09T08:13:36.378Z</atom:updated>
            <content:encoded><![CDATA[<p>We’ve all been there. You start a new Oracle APEX project, and before you can build the “cool” features, you have to spend hours designing a user table, a roles table, and the logic to connect them.</p><p>For my recent project, I didn’t want that overhead. I needed a way to separate <strong>Students</strong> from <strong>Admins</strong> instantly. Then it hit me: I already had the answer in the user’s email address.</p><p>Here is how I used a simple “Email Trick” to automate my app security.</p><h3>The Logic: Who are they?</h3><p>In my app, I have two types of users:</p><ol><li><strong>Students:</strong> They always log in using their personal <strong>Google accounts</strong> (@gmail.com).</li><li><strong>Admins:</strong> They log in using <strong>internal APEX accounts</strong> (which don’t use Gmail).</li></ol><p>Instead of checking a database table, I decided to check the <strong>string</strong> of the email address itself.</p><h3>Step 1: Creating a Global “Flag”</h3><p>I created an <strong>Application Item</strong> called LOGIN_USER_ROLE. Think of this as a temporary tag that follows the user around while they are using the app.</p><ul><li>Got to Application &gt; <strong>Shared Components</strong> &gt;<strong>Application Items.</strong></li><li>Click <strong>Create</strong>.<br>- Name: <strong>LOGIN_USER_ROLE</strong><br>- Scope: <strong>Application</strong></li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*sSqzW2RrcEjxZ9W5L2YO2w.png" /><figcaption>Application Item</figcaption></figure><h3>Step 2: The Automated Gatekeeper</h3><p>On my Home Page (Page 1), I added a <strong>Before Header</strong> process. The goal is simple: the moment the page starts to load, the app looks at the username and decides their role.</p><ul><li>Go to <strong>Home Page</strong> (page 1)</li><li>create a <strong>New Process</strong> in Prerendering Before Header Section</li><li>Name it <strong>SetUserRole</strong></li><li>use below <strong>PL/SQL</strong> Code.</li></ul><pre>begin<br>    if :APP_USER like &#39;%@gmail.com&#39; then<br>        :LOGIN_USER_ROLE := &#39;STUDENT&#39;;<br>    else<br>        :LOGIN_USER_ROLE := &#39;ADMIN&#39;;<br>    end if;<br>end;</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jUVi1v9mvSz9xNOyJ7E7pA.png" /><figcaption>Before Header Process.</figcaption></figure><h3>Step 3: Setting Up the “Security Guards”</h3><p>I created two <strong>Authorization Schemes</strong> in Shared Components:</p><ul><li><strong>IS_ADMIN:</strong> Returns TRUE only if :LOGIN_USER_ROLE is &#39;ADMIN&#39;.</li><li>Got to Application &gt; <strong>Shared Components</strong> &gt;<strong>Authorization scheme.</strong></li><li>Click <strong>Create</strong>.<br>- Name: <strong>IS_ADMIN<br>- </strong>Scheme Type: <strong>PL/SQL Function Returning Boolean</strong> <br>- PL/SQL Function Body(Value Required) :</li></ul><pre>if :Login_User_Role =&#39;ADMIN&#39; THEN<br>RETURN TRUE;<br>ELSE<br>RETURN FALSE;<br>END IF;</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*0_59fzlWZj-sY6FWMHeePA.png" /><figcaption>Authorization Scheme</figcaption></figure><ul><li><strong>IS_STUDENT:</strong> Returns TRUE only if :LOGIN_USER_ROLE is &#39;STUDENT&#39;.</li><li>Got to Application &gt; <strong>Shared Components</strong> &gt;<strong>Authorization scheme.</strong></li><li>Click <strong>Create</strong>.<br>- Name: <strong>IS_STUDENT<br>- </strong>Scheme Type: <strong>PL/SQL Function Returning Boolean</strong> <br>- PL/SQL Function Body(Value Required) :</li></ul><pre>if :Login_User_Role =&#39;STUDENT&#39; THEN<br>RETURN TRUE;<br>ELSE<br>RETURN FALSE;<br>END IF;</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*JoVHYiJd6iuEx_ZzLS_AWA.png" /><figcaption>Authorization Scheme</figcaption></figure><h3>Step 4: Locking the Doors (Pages &amp; Navigation)</h3><p>This is the best part. I went to my <strong>Navigation Menu</strong> and the <strong>Page Properties</strong> and simply applied these schemes.</p><ul><li>If a student logs in, the <strong>Admin pages </strong>link simply vanishes from the menu.</li><li>If an Admin logs in, but they don’t see the student-specific pages.</li></ul><h3>Apply Scheme to pages:</h3><ul><li>Got to Application &gt; P<strong>age 2: Student Page.</strong></li><li>Click on the Page Name &gt; <strong>Security</strong>&gt;<strong>Authorization Scheme.</strong></li><li>Select : <strong>IS_STUDENT</strong></li></ul><blockquote>Same way do for <strong>Page 3</strong>: Admin Page, set Scheme to <strong>IS_ADMIN.</strong></blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4j7EV0HeNuKM-p_Z7uoRng.png" /><figcaption>Student Page</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*nmpEZs88Gjlqx-jOdJNUAw.png" /><figcaption>Admin Page</figcaption></figure><h3>Apply Scheme to Navigation:</h3><ul><li>Got to Application &gt; <strong>Shared Components</strong> &gt; <strong>List </strong>&gt;<strong>Navigation scheme.</strong></li><li>set the appropriate scheme to navigation pages.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*eLM7zQoZ79SDNppSSXW-PQ.png" /><figcaption>Navigation Scheme.</figcaption></figure><h3>Working Application:</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*c_-DWuE5YtKnrW9gHyp3KA.gif" /></figure><h3>Why I Love This Approach</h3><ol><li><strong>Zero Maintenance:</strong> I don’t have to add new users to a “Roles” table. If they have a Gmail address, the app already knows what to do.</li><li><strong>Speed:</strong> I built the entire security layer in about 5 minutes.</li><li><strong>Low Code:</strong> It uses native APEX features with just a tiny bit of PL/SQL.</li></ol><h3>A Quick Tip for Developers</h3><p>If you try this, remember that I put my logic on the <strong>Home Page</strong>. If you want to be even safer, you can move that logic to a <strong>Post-Authentication Procedure</strong> in your Authentication Scheme. This ensures the role is set the very millisecond they log in, no matter which page they land on first!</p><h3>Final Thoughts</h3><p>Sometimes the best solution isn’t the most complex one. By looking at the data I already had (the email), I saved myself from managing unnecessary tables.</p><p><strong>Are you a fan of “Zero-Table” logic, or do you prefer the traditional RBAC approach? Let me know in the comments!</strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c4c4862cdf20" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Stop Managing Passwords: How I Integrated Google Login into my Oracle APEX App]]></title>
            <link>https://medium.com/@sarfarazMerchant/stop-managing-passwords-how-i-integrated-google-login-into-my-oracle-apex-app-45921bc5cb08?source=rss-5258823104b4------2</link>
            <guid isPermaLink="false">https://medium.com/p/45921bc5cb08</guid>
            <category><![CDATA[oracle]]></category>
            <category><![CDATA[google-authentication]]></category>
            <category><![CDATA[oracle-apex]]></category>
            <category><![CDATA[ace-oracle]]></category>
            <dc:creator><![CDATA[Sarfaraz Merchant]]></dc:creator>
            <pubDate>Fri, 08 May 2026 04:55:55 GMT</pubDate>
            <atom:updated>2026-05-08T04:55:55.689Z</atom:updated>
            <content:encoded><![CDATA[<p>Managing a user table with encrypted passwords is a headache I wanted to avoid for my latest student management project. I wanted a “One-Tap” experience where students could just use their existing Google accounts.</p><p>In this post, I’ll walk you through exactly how I set up <strong>Google Social Sign-In</strong> in Oracle APEX. No complex PL/SQL packages or custom handshakes — just standard APEX features and a bit of Google Cloud configuration.</p><h3><strong>Step 1: The Google Cloud</strong></h3><p>First, you need to tell Google that your APEX app is allowed to talk to it.</p><ol><li>Head over to the <a href="https://console.cloud.google.com/"><strong>Google Cloud Console</strong></a>.</li><li>Create a new project (call it something like My-APEX-App).</li><li>Go to <strong>APIs &amp; Services &gt; Credentials</strong></li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*SfDOxo0rLmVDcf4n86KiCw.png" /></figure><p>4. Click <strong>Create Credentials &gt;</strong>C<strong>onfigure consent screen</strong>.</p><blockquote>To create an OAuth client ID, we must first configure your consent screen. Click C<strong>onfigure consent screen.</strong></blockquote><p>=&gt; Fill Below details in Project configuration.</p><ul><li><strong>App Information<br>- App Name: </strong>Apex Google Sign In<br>- <strong>User Support Email</strong>: your email Address</li><li><strong>Audience</strong>: External</li><li><strong>Email Address</strong>: your email Address</li><li>Click <strong>Create.</strong></li></ul><p>5. Click Again &gt; Create Credentials&gt; <strong>OAuth client ID</strong>.</p><p>=&gt; Details</p><ul><li>Application Type: Web Application</li><li>Name: Google Sign in Application</li><li>Authorized redirect URIs &gt; Click <strong>+Add URI</strong></li></ul><blockquote><strong>Format</strong>: <a href="https://[YOUR_APEX_INSTANCE_URL]/ords/apex_authentication.callback">https://[YOUR_APEX_INSTANCE_URL]/ords/apex_authentication.callback</a></blockquote><blockquote><strong>Example Sample</strong>: <a href="https://asflkj97890sdfkh-sdf987sdfkljls3456.adb.ap-mumbai-1.oraclecloudapps.com/ords/apex_authentication.callback">https://asflkj97890sdfkh-sdf987sdfkljls3456.adb.ap-mumbai-1.oraclecloudapps.com/ords/apex_authentication.callback</a></blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Q5rdOb2GLjuNZLN5Q-2JQw.png" /><figcaption>OAuth Client ID.</figcaption></figure><ul><li><em>Tip: If you are on apex.oracle.com, use their specific callback URL.</em></li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/488/1*Ab1V6d90MNwL6jpMLV4MhQ.png" /><figcaption>Oauth Created.</figcaption></figure><blockquote>You will no longer be able to view or download the client secret once you close this dialog. Make sure you have copied or downloaded the information below and securely stored it.</blockquote><h3>Step 2: Storing Credentials Safely in APEX</h3><p>We don’t want to paste secrets directly into our code. APEX has a dedicated place for this.</p><ul><li>Open <strong>App Builder</strong>.</li><li>Navigate to <strong>Workspace Utilities &gt; Web Credentials</strong>.</li><li>Click <strong>Create</strong>.</li><li>Give it a Name.</li><li>Authentication Type : Basic</li><li>Enter your Client ID and Client Secret from Step 1.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*0mCw22s1Ro8UOlyQNPi8rg.png" /><figcaption>Web Credentials.</figcaption></figure><h3>Step 3: Setting up the Authentication Scheme</h3><p>Now, let’s tell the app to actually use Google.</p><ol><li>Create a Sample Apex Application</li><li>Go to <strong>Shared Components &gt; Authentication Schemes</strong>.</li><li>Click <strong>Create &gt; Based on a Pre-Configured Scheme from the Gallery</strong>.</li></ol><ul><li>Name : <strong>Google Authentication</strong></li><li>Scheme Type: <strong>Social Sign-In</strong></li><li>Credential Store: <strong>Google Sign In</strong></li><li>Authentication Provider: <strong>Google</strong></li><li>Scope: <strong>profile,email</strong></li><li>Username Interface<strong>:</strong> <strong>#email#.</strong></li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*JiCMqSiCMv1BMKtiw6B7iw.png" /><figcaption>Authentication Scheme.</figcaption></figure><p>4. Scroll Down and look for Login Processing and toggel Swith in Session to <strong>Enabled.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*klKajzvPwTUxf-zP9n3HQw.png" /><figcaption>Switch in Session.</figcaption></figure><h3><strong>Step 4: Add Google Login Button</strong></h3><p>We Will Add a Google Login Button in Login page.</p><ul><li>Open Your Application.</li><li>Go to <strong>Login page</strong> (9999-Login Page).</li><li>Create a <strong>New Button </strong>Below Login Button.<br>- Name : <strong>Google</strong><br>- Label : <strong>Sign in With Google</strong><br><strong>Appereance </strong>: <br>- Hot: <strong>On</strong><br>- Icon : fa-google<br><strong>Behavior:<br>- </strong>Action <strong>: Redicrect to page in application<br>Tempalte:<br>- </strong>Target<strong>: Page 1<br>- Advanced<br>- Request:</strong>APEX_AUTHENTICATION=Google Authentication.</li></ul><blockquote>Google Authentication is our authentication Scheme</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*kfE5zonI7_daJ0gBKqbkXA.png" /><figcaption>Button Design.</figcaption></figure><p>Here is the Working Application:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*gyoYSEGhyLkithDXO1Zx1w.gif" /></figure><h3>Wrapping Up</h3><p>Setting up Google Authentication shouldn’t take days of coding. By using the native <strong>Social Sign-In</strong> features in Oracle APEX, we’ve secured our app with enterprise-grade MFA and simplified the user experience in under 30 minutes.</p><p><strong>A quick reminder:</strong> If you find yourself stuck in a redirect loop after clicking “Login,” double-check your <strong>Redirect URI</strong> in the Google Console and make sure <strong>“Session Switch”</strong> is enabled in your APEX scheme. That small toggle saved my project!</p><p>Now that our “front door” is open and students can log in securely, we have a new challenge: <strong>How do we make sure a student doesn’t accidentally access the Admin dashboard?</strong></p><p>In my next post, I’ll show you how to build a robust <strong>Role-Based Authorization</strong> system to keep your data safe.</p><p><em>Are you using Social Sign-In for your internal apps, or do you prefer traditional LDAP? Let me know in the comments!</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=45921bc5cb08" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Building a Custom AI Assistant Inside Your Oracle APEX App]]></title>
            <link>https://pub.towardsai.net/building-a-custom-ai-assistant-inside-your-oracle-apex-app-26b229679e93?source=rss-5258823104b4------2</link>
            <guid isPermaLink="false">https://medium.com/p/26b229679e93</guid>
            <category><![CDATA[oracle-database]]></category>
            <category><![CDATA[ai-agent]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[oracle-apex]]></category>
            <category><![CDATA[ai-oracle]]></category>
            <dc:creator><![CDATA[Sarfaraz Merchant]]></dc:creator>
            <pubDate>Wed, 08 Apr 2026 10:28:21 GMT</pubDate>
            <atom:updated>2026-04-09T04:26:36.099Z</atom:updated>
            <content:encoded><![CDATA[<blockquote><em>In</em><strong><em> Part 1</em></strong><em> of this series </em><a href="https://medium.com/@sarfarazMerchant/getting-started-with-ai-for-free-a-step-by-step-guide-to-your-openrouter-api-key-390ad7b43fa6"><strong><em>Guide to Your OpenRouter API Key</em></strong></a><em>, we grabbed our Free “Universal API Key” from </em><strong><em>OpenRouter</em></strong><em>.</em></blockquote><blockquote>In <strong>Part 2</strong>, we used it to architect a full <a href="https://medium.com/@sarfarazMerchant/building-your-first-ai-app-in-oracle-apex-for-free-6840da33215b"><strong>HRMS Portal from scratch using AI</strong></a>.</blockquote><p><strong>What if your database could actually talk back to you?</strong> Imagine a world where you don’t need a report just to find an employee’s remaining leave balance or their last salary hike. What if you could convert your entire data warehouse into a conversational partner, getting instant answers by simply asking questions in <strong>Natural Language</strong>?</p><blockquote>Today, in Part 3 of this series we are going to give that application a “brain.”</blockquote><p>Most enterprise apps are great at showing data, but they struggle to answer simple questions like, <em>“Who is the highest-paid employee in IT?”</em> or <em>“How many leave requests are pending?”</em> Today, we will build an <strong>AI Assistant</strong> inside APEX to do exactly that.</p><h3>Prerequisite:</h3><ul><li>Must Have API Key from Previous Blog Part 1.</li><li>Must be Having an APEX instance to Practice this demo. (if you are new to Oracle Apex then create your free workspace now :<a href="https://apex.oracle.com/en/learn/getting-started/">Get started with APEX.</a></li><li>You should have a Sample Application, as we have created in Part 2</li></ul><h3>Step 1: Create Generative AI Configuration</h3><p>First, we need to establish the connection between APEX and the AI model.</p><ul><li><strong>Navigation:</strong> Go to your application in <strong>App Builder</strong> -&gt; <strong>Shared Components</strong> -&gt; <strong>Generative AI-&gt;</strong> <strong>AI Configurations</strong> -&gt; <strong>Create</strong>.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*HWozAEVOmZK_k4gK8iWhlA.png" /><figcaption>AI Configurations</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Ov_b1gHYjF5JfgAHiE9sEQ.png" /><figcaption>Create New Configuration</figcaption></figure><ul><li><strong>Name:</strong> Employee APP AI Configurations</li><li><strong>Service:</strong> OpenRouter_Free_AI (Service We Created in Part 2 of the Series)</li><li><strong>System Prompt:</strong> &gt; “You are an AI Assistant. Help answer queries related to employees, their departments, salary details, and leave status based on the provided data.”</li><li><strong>Welcome Message:</strong> “Hi, how may I help you today?”</li><li>Click <strong>Create </strong>Button.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*AJFuKTlg1QXuYxfjE3ETcg.png" /><figcaption>Configuration Details</figcaption></figure><blockquote>As soon as you click create button you will be redirected to <strong>RAG Sources.</strong></blockquote><h3>Step 2: Create RAG Source</h3><p>Now, we give the AI “eyes” to see your data by creating a RAG source using a specialized SQL query.</p><ul><li><strong>Navigation:</strong> Within your <strong>Employee APP AI Configurations</strong>, click on the <strong>RAG Sources</strong> tab -&gt; <strong>Create RAG Source</strong>.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*nXojMotHEoYAX6pSliHxJw.png" /><figcaption><strong>Create RAG Source</strong></figcaption></figure><ul><li><strong>Name:</strong> Employee Assistant RAG</li><li><strong>Description:</strong> Fetches comprehensive employee, department, salary, and leave details.</li><li><strong>Source Type:</strong> SQL Query</li><li><strong>SQL Query:</strong></li></ul><blockquote>SELECT <br> e.id AS employee_id,<br> e.first_name || ‘ ‘ || e.last_name AS full_name,<br> e.email,<br> e.hire_date,<br> e.job_id,<br> d.name AS department_name,<br> d.location AS department_location,<br> m.first_name || ‘ ‘ || m.last_name AS manager_name,<br> sh.salary_amount AS current_salary,<br> lr.leave_type,<br> lr.status AS leave_status<br>FROM hr_employees_1 e<br>JOIN hr_departments_1 d ON e.department_id = d.id<br>LEFT JOIN hr_employees_1 m ON e.manager_id = m.id<br>LEFT JOIN (<br> SELECT employee_id, salary_amount <br> FROM (SELECT employee_id, salary_amount, <br> ROW_NUMBER() OVER (PARTITION BY employee_id ORDER BY effective_date DESC) as rn <br> FROM hr_salary_history_1) WHERE rn = 1<br>) sh ON e.id = sh.employee_id<br>LEFT JOIN (<br> SELECT employee_id, leave_type, status <br> FROM (SELECT employee_id, leave_type, status, <br> ROW_NUMBER() OVER (PARTITION BY employee_id ORDER BY updated_on DESC) as rn <br> FROM hr_leave_requests_1) WHERE rn = 1<br>) lr ON e.id = lr.employee_id;</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jWmKb-XcHs0dp9buIptBvw.png" /><figcaption>Configure RAG Source</figcaption></figure><h3>Step 3: Create AI Assistant Chat Interface</h3><p>Finally, we add the UI component to your application so users can actually interact with the bot.</p><h4>3.1) Create the Launch Button</h4><p>Navigate to the page where you want the chatbot (e.g., the Employee Dashboard).</p><ul><li>Right-click on <strong>Breadcrumb Bar</strong> and click <strong>Create Button</strong>.</li><li><strong>Name:</strong> AIAssistant</li><li><strong>Region:</strong> Breadcrumb</li><li><strong>Button Template:</strong> Text with Icon</li><li><strong>Icon:</strong> fa-ai-microchip</li><li><strong>Behavior Action:</strong> Defined by Dynamic Action</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*P3mzeGLuynpxkeXURcPk4A.png" /><figcaption>Adding Buttoon to Chat interface in Employee Page.</figcaption></figure><h4>3.2) Create the Dynamic Action</h4><ul><li>Right-click on the new button -&gt; <strong>Create Dynamic Action</strong>.</li><li><strong>Name:</strong> Assistant Chatbot</li><li><strong>True Action:</strong> Show AI Assistant</li><li><strong>Configuration:</strong> Employee APP AI Configurations</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*n2KzpICjcu5Y241qakGKZg.png" /><figcaption>Dynamic Action</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Pu6yzAxm-z8n5zA91zHfsw.png" /><figcaption>True Action to open Chat interface</figcaption></figure><h4>Result: Live Demo</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*AE2XMhkzdED3n4WXGiMiVg.gif" /></figure><h3>Conclusion: What We Achieved</h3><p>We successfully transformed static HR data into a <strong>conversational interface</strong>.</p><ul><li><strong>Live Insights:</strong> The bot uses RAG to fetch real-time salary and leave status.</li><li><strong>Efficiency:</strong> Users get answers via natural language instead of running manual reports.</li><li><strong>Native Integration:</strong> A modern, AI-first experience built entirely within the APEX framework.</li></ul><h3>What’s Next?</h3><p>This concludes our 3-part series on building AI-powered applications in Oracle APEX. But I am not stopping here! Stay tuned for my upcoming deep dives into the Oracle AI ecosystem, including:</p><ul><li><strong>AI in Oracle Integration Cloud (OIC)</strong></li><li><strong>AI in Oracle Fusion</strong></li></ul><h3>I Want to Hear From You!</h3><p>Would you like to see a <strong>video tutorial</strong> covering this entire 3-part blog series? I’d love to walk you through the setup live!</p><p><strong>Let me know in the comments if a video version would be helpful for your learning journey!</strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=26b229679e93" width="1" height="1" alt=""><hr><p><a href="https://pub.towardsai.net/building-a-custom-ai-assistant-inside-your-oracle-apex-app-26b229679e93">Building a Custom AI Assistant Inside Your Oracle APEX App</a> was originally published in <a href="https://pub.towardsai.net/">Towards AI</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Building Your First AI App in Oracle APEX — For Free]]></title>
            <link>https://pub.towardsai.net/building-your-first-ai-app-in-oracle-apex-for-free-6840da33215b?source=rss-5258823104b4------2</link>
            <guid isPermaLink="false">https://medium.com/p/6840da33215b</guid>
            <category><![CDATA[tech]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[oracle]]></category>
            <category><![CDATA[ace-oracle]]></category>
            <category><![CDATA[oracle-apex]]></category>
            <dc:creator><![CDATA[Sarfaraz Merchant]]></dc:creator>
            <pubDate>Tue, 24 Feb 2026 05:48:49 GMT</pubDate>
            <atom:updated>2026-04-09T04:48:15.318Z</atom:updated>
            <content:encoded><![CDATA[<blockquote>In Part 1 of this series <a href="https://medium.com/@sarfarazMerchant/getting-started-with-ai-for-free-a-step-by-step-guide-to-your-openrouter-api-key-390ad7b43fa6"><strong>Guide to Your OpenRouter API Key</strong></a>, we grabbed our “Universal Key” from <strong>OpenRouter</strong>. Today, we are going to put that key to work.</blockquote><h3>Prerequisite:</h3><ul><li>Must Have API Key from Previous Blog.</li><li>Must be Having an APEX instance to Practice this demo. (if you are new to Oracle Apex then create your free workspace now :<a href="https://apex.oracle.com/en/learn/getting-started/">Get started with APEX.</a></li></ul><h3>our Target:</h3><p>we will create an HRMS Application which will be having a Dashboard with charts and 4 different types of pages where we can manage the Employees, department and its salary.</p><p>If you have explored the new <strong>Generative AI Services</strong> in Oracle APEX, you probably noticed that the default options (OCI genrative AI Service, Cohere, and OpenAI) usually require a paid subscription or a complex cloud setup.</p><p><strong>Here is the secret:</strong> Since OpenRouter uses an “OpenAI-compatible” API format, we can use a <strong>Free AI Model.</strong></p><h3>Why this matters:</h3><ul><li><strong>Zero Cost:</strong> No credit card required.</li><li><strong>Standard Learning:</strong> You learn the <em>exact</em> same skills as if you were using paid OCI or OpenAI services.</li><li><strong>Flexibility:</strong> You can switch between Llama 3, Gemini, or Mistral without changing your APEX settings.</li></ul><h3>Step 1: Configure Free AI Model in Oracle APEX</h3><p>Before we build an app, we need to tell APEX where to send its AI requests.</p><ol><li><strong>Workspace Utilities:</strong> Log in to your workspace and navigate to <strong>App Builder &gt; Workspace Utilities &gt; Generative AI</strong>.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*wlGrEfVKYw2YYG5OtJ2piQ.png" /><figcaption>App Builder</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*r2tEZ3jGYWkHuIEdlt8pYQ.png" /><figcaption>Workspace Utilities</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*kKtItdptXNv4FjsYZqGBdA.png" /><figcaption>Genrative AI</figcaption></figure><p><strong>2. Create Service:</strong> Click <strong>Create</strong>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*GkLZnrQaMrwwwGfmryxQEQ.png" /><figcaption>Create AI Service</figcaption></figure><p><strong>3. Select Provider:</strong> Choose <strong>OpenAI</strong>. (Don’t worry, we are going to point it to OpenRouter).</p><p>4. Configuration:-<strong>:</strong></p><ul><li>Name<strong>:</strong> <strong>OpenRouter_Free_AI</strong> (<em>You can Name ANythin</em>g)</li><li>Base URL<strong>: </strong><a href="https://openrouter.ai/api/v1"><strong>https://openrouter.ai/api/v1</strong></a></li><li>API Key<strong>:</strong> Paste the key you created in <a href="https://medium.com/@sarfarazMerchant/getting-started-with-ai-for-free-a-step-by-step-guide-to-your-openrouter-api-key-390ad7b43fa6"><strong>Blog #1</strong></a><strong>.</strong></li><li>AI Model<strong>:</strong> In the “Model” field, type the name of a free model from OpenRouter, such as “<strong><em>openai/gpt-oss-120b:free”</em></strong>.</li><li>Click on Test Connection, you will see a <strong>connection sucessfull</strong> Message.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*zIpyns0MU7_Ayc-KuonApw.png" /><figcaption>OpenAI Free AI model</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*wcjgVM0SUqyXw6iw5Fe0AA.png" /><figcaption>Genrative AI Service Configuration.</figcaption></figure><blockquote><strong>Note</strong>: <em>Toggel the </em><strong>Use By App Builde</strong>r<em> button</em></blockquote><h3>Step 2: Create Your AI-Powered Data Model for Application</h3><p>Now for the magic. We are going to let APEX build an Data Model for us using the service we just created.</p><p>Assume that we want to create basic HRMS System Application, so we will prompt AI to generate a Data Model with Sample data for us to have a backbone of our Application.</p><ol><li><strong>Create Data Model :</strong> Go to <strong>SQL WorkShop&gt; Create Data Model Using AI</strong>.</li></ol><blockquote>Sample Prompt:<br>“Act as a Senior Database Architect. Design a clean, normalized <strong>HRMS Data Model</strong> using the prefix <strong>HR_</strong>.</blockquote><blockquote><strong>Include the following modules and table structures:</strong></blockquote><blockquote><strong>Core HR:</strong> Tables for <strong>DEPARTMENTS</strong> (ID, Name, Location) and <strong>EMPLOYEES</strong> (ID, First_Name, Last_Name, Email, Hire_Date, Job_ID, and Manager_ID).</blockquote><blockquote><strong>Leave Management:</strong> A table for <strong>LEAVE_REQUESTS</strong> linked to Employees with columns for Request_ID, Type (Sick, Vacation, Personal), Start_Date, End_Date, and Status (Pending, Approved, Rejected).</blockquote><blockquote><strong>Payroll:</strong> A table for <strong>SALARY_HISTORY</strong> to track Employee_ID, Effective_Date, and Salary_Amount.</blockquote><blockquote><strong>Requirements:</strong></blockquote><blockquote>Generate <strong>10 rows of realistic sample data</strong> for every table.</blockquote><blockquote>Include <strong>Audit Columns</strong> for every table: Created_By (default to APP_USER) and Created_Date (default to local timestamp).</blockquote><blockquote>Ensure all Primary Keys and Foreign Keys are explicitly defined.</blockquote><blockquote>Output the result as a <strong>ready-to-run Oracle SQL Script</strong>.”</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/876/1*FsXEJE8QBS6vZpP-twdKlg.png" /></figure><p><strong>2. Click on </strong><em>Create SQl Script</em><strong> and in Script Editor Name the Script and Click on </strong><em>Run </em><strong>Button.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*QYfVnkjCTNg8O9rQ1LPl6A.png" /><figcaption>Script Editor.</figcaption></figure><blockquote>Note: there is a possibility of getting Errors While Running this Script, i faced issues of “<strong>unknown statements</strong>”.<br>The Best way to resolve this is to provide this SQL Script to any of the AI Models (Gemini,ChatGPT,Claude), get the corrected version and try it will get Execute sucessfully.</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jOWVHzL9BhNDXVAT1_bbIg.png" /><figcaption>Script Result.</figcaption></figure><p>3. Click on <strong><em>Create App </em></strong>button then <strong><em>Create Application</em></strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/791/1*Twpq3zu9aqBwHIBgQ6n9zQ.png" /><figcaption>Create Application.</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*atYU3VkTaNzh3qEyiim9bA.png" /></figure><blockquote>This Step is required to make an entry of newly created tables in metadeta. (this Application can be deleted on later stage)</blockquote><p>Go Back to App Builder as Now we will be creating the Application Using AI.</p><h3>Step 3: Create Application Using Genrative AI</h3><ol><li>Go to App Builde-&gt;Create-&gt; Create Application using Genrative AI.</li></ol><p><strong>Prompt the AI:</strong> Tell the assistant what you want to build.</p><blockquote><em>“</em>Build a professional HRMS Portal using my existing tables: HR_DEPARTMENTS_1, HR_EMPLOYEES_1, HR_LEAVE_REQUESTS_1, and HR_SALARY_HISTORY_1.</blockquote><blockquote>Key Requirements:</blockquote><blockquote><strong>Executive Dashboard</strong>: Create a primary dashboard with a Pie Chart showing ‘Employees per Department’ and a Bar Chart showing ‘Leave Requests by Status’.</blockquote><blockquote><strong>Staff Directory:</strong> Create an Interactive Report for HR_EMPLOYEES_1 with a Faceted Search to filter by Department and Hire Date.</blockquote><blockquote><strong>Leave Management</strong>: Add a Calendar View based on HR_LEAVE_REQUESTS_1 using the Start_Date and End_Date columns.</blockquote><blockquote><strong>Payroll Insight</strong>: Create an Interactive Grid for HR_SALARY_HISTORY_1 so I can view and update salary records quickly.</blockquote><blockquote><strong>Forms</strong>: Ensure every report has a modal form for adding or editing records in the existing tables.</blockquote><blockquote>Please use the existing columns and relationships to build the application logic.<em>”</em></blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/813/1*WYhFHq2acHhZSiM00IlSvA.png" /><figcaption>Promt to Assistant</figcaption></figure><p>2. Click on <strong><em>Create Application</em></strong> and you will see a window where the Name is given Automatically and the Pages created as we have prompted.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*8523kLm7nWP_ZRUhNOGDNA.png" /><figcaption>Application creation Window</figcaption></figure><p>3. Click on <strong>Run Application</strong> to see the Application in Action.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*BR-ynGfWrvvUuG-AN-5A4Q.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/617/1*udVy6Ir3koOrv7rt-aEpKQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*J7yBGFa5jJnbgy6y15xsfQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DqJRDq_XMV-fbi8bKDhpCA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*K1Z-qZfHN-W9cTRpnZks8Q.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*VyrKS-u8gYXz2crG4GBsew.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*s3FrPSEu8WT65bQOONeImg.png" /></figure><h3><strong>Conclusion</strong></h3><p>Building this HRMS portal from scratch proved that with the right data model and a ‘Universal AI Key,’ the possibilities in Oracle APEX are endless.</p><p>What are you building with AI in APEX? I’d love to hear about your challenges or successes in the comments below. If you’re stuck on a specific SQL or AI configuration, let’s solve it together!</p><p>If this helped you, please share it with your network!</p><p><strong>Next Up:</strong> In Part 3, we’ll move beyond standard pages and build a custom AI Assistant that can actually ‘talk’ to this HRMS data.”</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6840da33215b" width="1" height="1" alt=""><hr><p><a href="https://pub.towardsai.net/building-your-first-ai-app-in-oracle-apex-for-free-6840da33215b">Building Your First AI App in Oracle APEX — For Free</a> was originally published in <a href="https://pub.towardsai.net/">Towards AI</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Getting Started with AI for Free: A Step-by-Step Guide to Your OpenRouter API Key]]></title>
            <link>https://medium.com/@sarfarazMerchant/getting-started-with-ai-for-free-a-step-by-step-guide-to-your-openrouter-api-key-390ad7b43fa6?source=rss-5258823104b4------2</link>
            <guid isPermaLink="false">https://medium.com/p/390ad7b43fa6</guid>
            <category><![CDATA[oracle-apex]]></category>
            <category><![CDATA[open-router]]></category>
            <category><![CDATA[free-ai-tools]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[api-key]]></category>
            <dc:creator><![CDATA[Sarfaraz Merchant]]></dc:creator>
            <pubDate>Mon, 16 Feb 2026 03:12:56 GMT</pubDate>
            <atom:updated>2026-02-16T03:12:56.751Z</atom:updated>
            <content:encoded><![CDATA[<p>In my 8+ years of working with Oracle technologies, I have realized that the biggest barrier to trying new tech isn’t the code — it’s the setup.</p><p>Right now, if you want to experiment with AI, you usually have to create five different accounts for five different AI models (like ChatGPT, Claude, or Gemini). It’s confusing for students and a management headache for professionals.</p><p>Today, I’m sharing a “shortcut” I use: <strong>OpenRouter</strong>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*aR0zYtwF5WeCkADcRaeSQA.png" /></figure><h3>What is OpenRouter?</h3><p>Think of it as a <strong>Universal Adapter</strong> for AI. Instead of getting 10 different keys for 10 different AI providers, you get <strong>one key</strong> that connects to almost every major AI model in existence.</p><h3>Why this is the best way to start in 2026:</h3><ul><li><strong>Zero Cost:</strong> You can access high-quality “Free” models without even adding a credit card.</li><li><strong>One Dashboard:</strong> You see all your usage in one place.</li><li><strong>Future-Proof:</strong> If a better AI model comes out tomorrow, you don’t need a new account. You just change one setting in your app.</li></ul><h3>Step-by-Step: Getting Your Key</h3><p>Whether you are a student building your first project or a professional architecting a new solution, here is how you get started in under 60 seconds:</p><ol><li><strong>Create Your Account</strong> Go to <a href="https://openrouter.ai/">OpenRouter.ai</a>. You can sign up instantly using your Google or GitHub account.</li></ol><p><strong>2. Navigate to “Keys”</strong> Click on your profile icon in the top right corner and select <strong>Keys</strong>.</p><p><strong>3. Generate Your Key</strong> Click the <strong>“Create Key”</strong> button. Give it a simple name like My_First_AI_Project.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*K7Xisd5a3q8NlB--Ukx0Yw.png" /><figcaption>List All API Key.</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*92Zsm5N4XeUotvdlxZTJTA.png" /><figcaption>Create API Key.</figcaption></figure><p><strong>4. Set a “Safety Switch” (Professional Tip)</strong> Even if you are using free models, I always recommend setting a “Credit Limit” (like $1 or $5). This is a great habit for <strong>Cost Governance</strong> — it ensures you never spend more than you intended.</p><p><strong>5. Secure the Key</strong> Click “Create” and <strong>copy the key immediately</strong>. Store it in a safe place (like a password manager or a private note). For security reasons, OpenRouter will never show this exact key to you again.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/715/1*_CYExYTxvzsrn7rSGcDK8g.png" /><figcaption>API Key.</figcaption></figure><h3>Finding Your “Free” Playground</h3><p>Now that you have your key, you need to know which AI models you can use for free. OpenRouter makes this very easy:</p><ol><li><strong>Browse the Models:</strong> Go to the <a href="https://openrouter.ai/models">OpenRouter Models</a> page.</li><li><strong>Filter for Free:</strong> In the search or filter bar, simply type <strong>“:free”</strong>.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*VocaeUll36DdZkTNFcYruw.png" /><figcaption>Models.</figcaption></figure><ul><li>This will show you a list of high-quality models (like <strong>OpenAI</strong>,<strong>Llama 3</strong>, <strong>Gemini etc</strong>) that cost $0 to use.</li></ul><ol><li><strong>Check the Docs:</strong> * Click on any model name (e.g., OpenAI: gpt-oss-120b (free)).</li></ol><ul><li>On that model’s page, you will see a <strong>“Quickstart”</strong> tab.</li><li>This tab gives you the exact “code snippet” you need to copy-paste into your application to start chatting with that specific AI.</li></ul><h3>What Can You Do With This?</h3><p>Now that you have your “Universal Key,” you are ready to build. You can plug this into Python scripts, web apps, or enterprise tools.</p><p>Personally, I use this key to power a custom <strong>AI Assistant inside Oracle APEX</strong>. It allows me to switch between different AI “brains” instantly to see which one gives my users the best answers.</p><h3>What’s Next?</h3><p>Now that we have our key, we need a place to use it.</p><p>In my next post, I will show you exactly how to take this key and configure it within an application (using <strong>Oracle APEX</strong> as our example). We will move from “having a key” to “having a working AI.”</p><p><strong>Are you a student or a pro? Let me know in the comments which AI model you’re planning to try first!</strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=390ad7b43fa6" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>