<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: GAURIKA SETHI</title>
    <description>The latest articles on DEV Community by GAURIKA SETHI (@gaurikasethi88).</description>
    <link>https://dev.to/gaurikasethi88</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4191136%2F59b3bf75-2124-4f9b-b09a-4d061af083d7.png</url>
      <title>DEV Community: GAURIKA SETHI</title>
      <link>https://dev.to/gaurikasethi88</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/gaurikasethi88"/>
    <language>en</language>
    <item>
      <title>Documents Have Ghosts. We Built an AI to Hunt Them</title>
      <dc:creator>GAURIKA SETHI</dc:creator>
      <pubDate>Wed, 28 Oct 2026 07:15:32 +0000</pubDate>
      <link>https://dev.to/gaurikasethi88/documents-have-ghosts-we-built-an-ai-to-hunt-them-2d27</link>
      <guid>https://dev.to/gaurikasethi88/documents-have-ghosts-we-built-an-ai-to-hunt-them-2d27</guid>
      <description>&lt;p&gt;&lt;em&gt;This is a submission for the &lt;a href="https://dev.to/challenges/hacktoberfest-week4-2026-10-26"&gt;Hacktoberfest Open-Source AI Challenge Week 4: Ghost in the Machine&lt;/a&gt;.&lt;/em&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  What I Built
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;What if the document you shared contained information you thought was already gone?&lt;/strong&gt;&lt;/p&gt;

&lt;p&gt;A document can look perfectly clean on the surface while still carrying sensitive information, forgotten comments, author metadata, tracked changes, or hidden text. We often assume that once something is deleted or edited, it disappears completely. Digital documents can tell a different story.&lt;/p&gt;

&lt;p&gt;That is the idea behind &lt;strong&gt;Ghost in the Document&lt;/strong&gt;, a local-first document privacy tool designed to uncover these hidden traces before a file is shared.&lt;/p&gt;

&lt;p&gt;The application scans supported PDF, DOCX, and TXT files for potentially sensitive information, examines supported document structures for selected hidden content, and lets users review the findings before choosing what to redact. It then creates a redacted copy and performs post-save checks to help verify the result.&lt;/p&gt;

&lt;p&gt;For this challenge, I interpreted &lt;em&gt;Ghost in the Machine&lt;/em&gt; as the remnants of information that remain inside our digital files long after we stop thinking about them.&lt;/p&gt;

&lt;p&gt;The goal was to turn that idea into something practical: a tool that helps people look beyond what is immediately visible and take greater control over the information they share, without needing to upload their documents to an online AI service.&lt;/p&gt;

&lt;h2&gt;
  
  
  Demo
&lt;/h2&gt;

&lt;p&gt;See Ghost in the Document in action, from scanning a document to reviewing findings and creating a redacted copy.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;&lt;a href="https://youtu.be/F676Gr8bNvs" rel="noopener noreferrer"&gt;Watch the demo on YouTube&lt;/a&gt;&lt;/strong&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  Code
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;GitHub Repository:&lt;/strong&gt; &lt;a href="https://github.com/Rishi11pixel/ghost-in-the-document" rel="noopener noreferrer"&gt;Ghost in the Document&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;The project is built with Python and brings together document processing, a Gradio interface, rule-based detection, open-source AI, and locally running language models.&lt;/p&gt;

&lt;p&gt;The code is available on GitHub for anyone interested in exploring the implementation or running the project locally.&lt;/p&gt;

&lt;h2&gt;
  
  
  How I Built It
&lt;/h2&gt;

&lt;p&gt;I did not want to rely on a single model to identify every possible privacy risk. Different kinds of information require different detection techniques, so I built the application around several complementary layers.&lt;/p&gt;

&lt;h3&gt;
  
  
  1. Rule-Based Detection
&lt;/h3&gt;

&lt;p&gt;The first layer uses regular expressions and predefined rules to identify recognizable patterns, such as email addresses and phone numbers.&lt;/p&gt;

&lt;p&gt;These checks are fast and predictable for information that follows known formats. They provide a useful first line of detection before more context-dependent analysis takes place.&lt;/p&gt;

&lt;h3&gt;
  
  
  2. GLiNER for Named Entity Recognition
&lt;/h3&gt;

&lt;p&gt;Not every sensitive detail follows a predictable pattern. Names, organizations, and locations can appear in many different forms.&lt;/p&gt;

&lt;p&gt;I integrated GLiNER to help identify these entities in document text, extending detection beyond what regular expressions alone can reasonably capture.&lt;/p&gt;

&lt;h3&gt;
  
  
  3. Ollama for Local Language Model Analysis
&lt;/h3&gt;

&lt;p&gt;Some information becomes sensitive because of its context rather than its format. A locally running language model through Ollama adds another layer of analysis to help identify contextual details that simpler methods might miss.&lt;/p&gt;

&lt;p&gt;I also included a safeguard that checks whether proposed findings appear verbatim in the source text. This helps reject invented text spans, although it does not eliminate every possible model error or guarantee complete detection.&lt;/p&gt;

&lt;h3&gt;
  
  
  4. Looking Beyond Visible Text
&lt;/h3&gt;

&lt;p&gt;The contents displayed when opening a document are only part of the picture.&lt;/p&gt;

&lt;p&gt;The application also examines supported document structures for selected hidden content, including metadata, comments, tracked changes, and hidden text. These checks help identify traces that may not be obvious to someone simply reading the document.&lt;/p&gt;

&lt;h3&gt;
  
  
  5. Review, Redaction, and Verification
&lt;/h3&gt;

&lt;p&gt;Finding a potential privacy issue is only the beginning. Users need control over what gets removed, and the application needs to check what happens afterward.&lt;/p&gt;

&lt;p&gt;Ghost in the Document lets users review findings and select the information they want to redact. It then creates a separate redacted copy and performs post-save checks on the output.&lt;/p&gt;

&lt;p&gt;I wanted this to be a review-driven workflow rather than a process that blindly removes everything a model flags. A successful scan or verification check should not be treated as proof that a document is completely free of sensitive information.&lt;/p&gt;

&lt;p&gt;The current version supports text-based PDFs, DOCX, and TXT files. Scanned PDFs and image-based documents requiring OCR are outside the current scope.&lt;/p&gt;

&lt;p&gt;One of the main challenges was bringing these different components together into a single, usable workflow. Detection, document processing, redaction, and verification each have their own limitations, and making them work together required treating them as separate but connected stages.&lt;/p&gt;

&lt;h2&gt;
  
  
  Why Does Open Innovation Matter?
&lt;/h2&gt;

&lt;p&gt;There is an obvious contradiction in uploading a private document to an external service just to check whether it contains private information.&lt;/p&gt;

&lt;p&gt;That is one of the main reasons I wanted to explore local AI.&lt;/p&gt;

&lt;p&gt;Open-source tools and locally runnable models make it possible to build a workflow where documents can be processed on the user's own machine, without depending on a paid, hosted AI API for the core local workflow.&lt;/p&gt;

&lt;p&gt;They also offer flexibility. Rule-based detection, named entity recognition, and language model analysis can be combined, tested, and improved independently. Developers can experiment with different models and approaches rather than being tied to a single closed service.&lt;/p&gt;

&lt;p&gt;For a privacy-focused application, that flexibility matters. It makes local processing more accessible and gives users greater control over how their information is handled.&lt;/p&gt;

&lt;p&gt;Of course, open-source AI and local execution do not automatically guarantee perfect privacy or accuracy. Models can miss sensitive information, produce false positives, and struggle with unusual document structures. Redaction and verification can also have limitations.&lt;/p&gt;

&lt;p&gt;That is why this project combines multiple detection methods with human review and post-save checks instead of treating any single model as the final authority.&lt;/p&gt;

&lt;h2&gt;
  
  
  The Spirit of It
&lt;/h2&gt;

&lt;p&gt;The unsettling thing about a ghost is not always that it exists. Sometimes, it is the fact that nobody knows it is there.&lt;/p&gt;

&lt;p&gt;Documents have a similar problem.&lt;/p&gt;

&lt;p&gt;A comment can survive long after a discussion has ended. A personal detail can remain buried in a paragraph. Information that seems to have disappeared can leave traces behind. We edit, save, and share files assuming their history has been left behind, but sometimes it travels with them.&lt;/p&gt;

&lt;p&gt;That became the central idea behind Ghost in the Document.&lt;/p&gt;

&lt;p&gt;I wanted to interpret the challenge theme in a way that was both creative and useful. Instead of building something spooky just for appearances, I treated the ghosts as the forgotten remnants of information that can follow our digital work.&lt;/p&gt;

&lt;p&gt;Building this project also reinforced an important lesson: privacy is not just about detecting sensitive information. It is about helping people understand what has been found, decide what should be removed, and check the result before sharing it.&lt;/p&gt;

&lt;p&gt;There is still room to improve detection coverage and strengthen verification, but the purpose remains simple.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Before a document leaves your machine, give it one last chance to reveal what it is carrying.&lt;/strong&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  Team
&lt;/h2&gt;

&lt;p&gt;Built by me, with my teammate &lt;a class="mentioned-user" href="https://dev.to/rishi_rajgoel_c5e99c8b81"&gt;@rishi_rajgoel_c5e99c8b81&lt;/a&gt;&lt;/p&gt;

</description>
      <category>devchallenge</category>
      <category>hf26challenge</category>
    </item>
    <item>
      <title>Recipe De-Blogger</title>
      <dc:creator>GAURIKA SETHI</dc:creator>
      <pubDate>Thu, 22 Oct 2026 16:41:59 +0000</pubDate>
      <link>https://dev.to/gaurikasethi88/recipe-de-blogger-pam</link>
      <guid>https://dev.to/gaurikasethi88/recipe-de-blogger-pam</guid>
      <description>&lt;p&gt;&lt;em&gt;This is a submission for the &lt;a href="https://dev.to/challenges/hacktoberfest-week3-2026-10-19"&gt;Hacktoberfest Open-Source AI Challenge Week 3: Slow and Steady&lt;/a&gt;&lt;/em&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  What I Built
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;Recipe De-Blogger&lt;/strong&gt; is a lightweight web app that helps home cooks get to the useful part of an online recipe without wading through long personal stories, ads, and extra page content.&lt;/p&gt;

&lt;p&gt;Paste a recipe URL and the app extracts the recipe into a cleaner, easier-to-follow format, including the title, ingredients, and cooking instructions. It can also scale ingredient quantities for a different serving size when the original serving count is available. The goal is simple: spend less time scrolling and more time cooking.&lt;/p&gt;

&lt;p&gt;The project is especially useful for people who cook from online recipes and want a quick, practical view they can follow while preparing a meal.&lt;/p&gt;

&lt;h2&gt;
  
  
  Demo
&lt;/h2&gt;

&lt;p&gt;&lt;a href="https://youtu.be/klun_BldR0c" rel="noopener noreferrer"&gt;Demo Video&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  Code
&lt;/h2&gt;

&lt;p&gt;&lt;a href="https://github.com/Rishi11pixel/recipe-deblogger.git" rel="noopener noreferrer"&gt;Repository Link&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;The project is built with a Python/FastAPI backend and a simple HTML, CSS, and JavaScript frontend. The backend extracts structured recipe data from recipe pages and uses a local language model as an AI-assisted fallback when needed.&lt;/p&gt;

&lt;h2&gt;
  
  
  How I Built It
&lt;/h2&gt;

&lt;p&gt;I built Recipe De-Blogger with a small, practical stack:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Python and FastAPI&lt;/strong&gt; power the backend API.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;HTML, CSS, and JavaScript&lt;/strong&gt; provide the frontend.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Recipe-page extraction&lt;/strong&gt; checks for structured recipe data, such as Schema.org recipe markup, before falling back to other extraction methods.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Ollama&lt;/strong&gt; runs an open-weight language model locally for AI-assisted recipe extraction, including extraction from available YouTube transcripts.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Recipe scaling logic&lt;/strong&gt; adjusts ingredient quantities for a target serving count when the original serving count is known.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;I focused on keeping the app straightforward: provide a URL, extract the recipe, show the important details, and make serving-size adjustments easier. Because extraction can vary between websites and transcript quality, the result should still be checked before cooking.&lt;/p&gt;

&lt;h2&gt;
  
  
  Why Does Open Innovation Matter?
&lt;/h2&gt;

&lt;p&gt;Open innovation made it possible to experiment with AI-assisted extraction without making every recipe request depend on a paid, closed AI API. Running a model locally through Ollama gives me more control over the development process and lets me learn how an open-weight model behaves on real, messy recipe content.&lt;/p&gt;

&lt;p&gt;It also keeps the architecture approachable. The app combines standard web technologies with a local model, so the extraction workflow can be inspected, tested, and improved as the project grows. For a small open-source project, that flexibility is valuable: I can iterate on prompts, parsing, and validation without tying the whole product to one proprietary model provider.&lt;/p&gt;

&lt;h2&gt;
  
  
  The Lazy Payoff
&lt;/h2&gt;

&lt;p&gt;The payoff is fewer unnecessary steps between finding a recipe and starting to cook. Instead of repeatedly scrolling past a blog's introduction and other page content, users can get the recipe details in one organized view.&lt;/p&gt;

&lt;p&gt;I haven't included a measured time-saving claim because I haven't run a formal timing test. The intended benefit is reducing friction: copy a recipe link, extract the key details, and focus on cooking.&lt;/p&gt;

&lt;h2&gt;
  
  
  Other Team Members
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;Rishi Raj Goel&lt;/strong&gt;- &lt;a class="mentioned-user" href="https://dev.to/rishi_rajgoel_c5e99c8b81"&gt;@rishi_rajgoel_c5e99c8b81&lt;/a&gt; &lt;/p&gt;

</description>
      <category>devchallenge</category>
      <category>hf26challenge</category>
    </item>
    <item>
      <title>Khata: Turning Paper Bills into Useful Business Records</title>
      <dc:creator>GAURIKA SETHI</dc:creator>
      <pubDate>Sun, 18 Oct 2026 16:52:44 +0000</pubDate>
      <link>https://dev.to/gaurikasethi88/khata-turning-paper-bills-into-useful-business-records-5c1e</link>
      <guid>https://dev.to/gaurikasethi88/khata-turning-paper-bills-into-useful-business-records-5c1e</guid>
      <description>&lt;h1&gt;
  
  
  Khata: Turning Paper Bills into Useful Business Records
&lt;/h1&gt;

&lt;p&gt;&lt;strong&gt;Submission:&lt;/strong&gt; Hacktoberfest Open-Source AI Challenge — Week 2: Side Hustle&lt;br&gt;&lt;br&gt;
&lt;strong&gt;Challenge:&lt;/strong&gt; &lt;a href="https://dev.to/challenges/hacktoberfest-week2-2026-10-12"&gt;https://dev.to/challenges/hacktoberfest-week2-2026-10-12&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  What I Built
&lt;/h2&gt;

&lt;p&gt;Small shops often handle bills and purchase records on paper. Entering those details into a ledger takes time, and handwritten or printed bills can be difficult to process consistently.&lt;/p&gt;

&lt;p&gt;I built &lt;strong&gt;Khata&lt;/strong&gt;, a local bill-to-ledger prototype for small shops in India. It uses AI to extract information from bill images and helps turn that information into structured records.&lt;/p&gt;

&lt;p&gt;The current prototype includes:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Bill image preprocessing using OpenCV.&lt;/li&gt;
&lt;li&gt;Local vision-based extraction using Ollama and Gemma 3 4B.&lt;/li&gt;
&lt;li&gt;Structured extraction output and validation checks.&lt;/li&gt;
&lt;li&gt;A Streamlit interface for reviewing and correcting extracted details.&lt;/li&gt;
&lt;li&gt;A local SQLite ledger for saved bills and line items.&lt;/li&gt;
&lt;li&gt;Monthly summaries and an experimental product-demand forecasting feature.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;The goal is to reduce repetitive data entry while keeping the shopkeeper in control of the final record.&lt;/p&gt;

&lt;h2&gt;
  
  
  Demo
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;Video demo:&lt;/strong&gt; &lt;a href="https://youtu.be/NfF9mBTL2Tg" rel="noopener noreferrer"&gt;Watch Khata Demo&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;In an initial test, one shopkeeper processed 10 bills:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Manual process: &lt;strong&gt;60 minutes&lt;/strong&gt;
&lt;/li&gt;
&lt;li&gt;Using Khata, including corrections: &lt;strong&gt;40 minutes&lt;/strong&gt;
&lt;/li&gt;
&lt;li&gt;Time saved: &lt;strong&gt;20 minutes (about 33%)&lt;/strong&gt;
&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;This is an early, small-scale test rather than a broad performance study.&lt;/p&gt;

&lt;h2&gt;
  
  
  Code
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;GitHub repository:&lt;/strong&gt; &lt;a href="https://github.com/Gaurika-Sethi/khata-" rel="noopener noreferrer"&gt;View Khata on GitHub&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  How I Built It
&lt;/h2&gt;

&lt;p&gt;Khata is built around a local-first workflow:&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt;A user uploads a bill image.&lt;/li&gt;
&lt;li&gt;OpenCV preprocessing improves the image for extraction.&lt;/li&gt;
&lt;li&gt;Ollama runs the Gemma 3 4B vision model locally to extract bill details.&lt;/li&gt;
&lt;li&gt;Pydantic-backed structure and validation checks help identify missing or inconsistent values.&lt;/li&gt;
&lt;li&gt;The user reviews and corrects the extracted information in the Streamlit interface.&lt;/li&gt;
&lt;li&gt;Confirmed records can be saved to a local SQLite database.&lt;/li&gt;
&lt;li&gt;The project includes scripts to build product-demand data and generate experimental restocking forecasts using TabPFN where there is enough historical data.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;The main tools are &lt;strong&gt;Python, Streamlit, OpenCV, Ollama, Gemma 3 4B, Pydantic, SQLite, pandas, and TabPFN&lt;/strong&gt;.&lt;/p&gt;

&lt;p&gt;I also evaluated extraction against a small set of 20 bills. The current results were:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Total accuracy: &lt;strong&gt;19/20 (95%)&lt;/strong&gt;
&lt;/li&gt;
&lt;li&gt;Item-count accuracy: &lt;strong&gt;17/20 (85%)&lt;/strong&gt;
&lt;/li&gt;
&lt;li&gt;Date accuracy: &lt;strong&gt;7/9 (77.8%)&lt;/strong&gt;, calculated only for the nine bills with usable ground-truth dates.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;These numbers describe this small test set and should not be treated as a guarantee of real-world accuracy. The prototype still needs testing on more shops, bill formats, and dates.&lt;/p&gt;

&lt;h2&gt;
  
  
  Why Does Open Innovation Matter?
&lt;/h2&gt;

&lt;p&gt;Small businesses should be able to experiment with useful technology without being locked into a single cloud service. Using an open-weight model locally gives developers more control over the workflow and makes it possible to inspect, adapt, and improve the system.&lt;/p&gt;

&lt;p&gt;Local inference can also reduce dependence on per-request API services, although the actual cost and privacy benefits depend on the hardware and deployment setup. Khata is an experiment in making practical AI tools more accessible while keeping the data workflow understandable and under the user's control.&lt;/p&gt;

&lt;h2&gt;
  
  
  My First User
&lt;/h2&gt;

&lt;p&gt;I tested Khata with one shopkeeper using 10 bills. The manual workflow took about 60 minutes, while using Khata—including corrections—took about 40 minutes. That is a saving of 20 minutes, or roughly 33%, in this initial test.&lt;/p&gt;

&lt;p&gt;The shopkeeper was &lt;strong&gt;maybe&lt;/strong&gt; willing to pay, depending on the features, speed, and reliability. Their expected price was around &lt;strong&gt;₹400–₹500 per month&lt;/strong&gt; if the tool becomes faster and consistently saves time.&lt;/p&gt;

&lt;p&gt;They also suggested two useful additions:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Tracking products or amounts given and taken on credit.&lt;/li&gt;
&lt;li&gt;End-of-day restocking alerts that account for sales and a delivery lead time of around 3–4 days.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;They mentioned Tally as an existing option, so accuracy and a clear advantage over current bookkeeping tools will matter if Khata is to be useful in real shops. This feedback is from one early user, so it is a starting point—not proof of product-market fit.&lt;/p&gt;

&lt;h2&gt;
  
  
  What I Learned
&lt;/h2&gt;

&lt;p&gt;Building Khata made one thing clear: extracting data is only one part of the problem. The output has to be easy to verify, errors need to be visible, and the resulting records must be useful for everyday decisions.&lt;/p&gt;

&lt;p&gt;The project is still a prototype. Its current limitations include inconsistent extraction on some bills, a small evaluation dataset, and limited historical data for demand forecasting. The next step is to improve reliability and test the workflow with more real bills and shopkeepers.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;What would make a tool like this genuinely useful for a small shop credit tracking, faster bill entry, restocking alerts, or something else?&lt;/strong&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  Other Team Members
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;Rishi Raj Goel&lt;/strong&gt; &lt;a class="mentioned-user" href="https://dev.to/rishi_rajgoel_c5e99c8b81"&gt;@rishi_rajgoel_c5e99c8b81&lt;/a&gt; &lt;/p&gt;

</description>
      <category>devchallenge</category>
      <category>hf26challenge</category>
    </item>
    <item>
      <title>New to community</title>
      <dc:creator>GAURIKA SETHI</dc:creator>
      <pubDate>Sat, 17 Oct 2026 20:44:27 +0000</pubDate>
      <link>https://dev.to/gaurikasethi88/new-to-community-5d28</link>
      <guid>https://dev.to/gaurikasethi88/new-to-community-5d28</guid>
      <description>&lt;p&gt;Hello everyone!!!&lt;/p&gt;

</description>
    </item>
  </channel>
</rss>
