diff --git a/README.md b/README.md
index 5067ca9..cac4ad6 100644
--- a/README.md
+++ b/README.md
@@ -74,7 +74,7 @@ flowchart LR
     app -- create embedding --> embeddings
     app -- search embeddings --> vdb
     app -- retrieve documents --> db
-    app -- get text completion --> llm
+    app -- fetch text completion --> llm
 ```
 
 The web application collects user input, creates an embedding, searches the vector database for similar embeddings,
diff --git a/src/app.ts b/src/app.ts
index 57731e5..374c051 100644
--- a/src/app.ts
+++ b/src/app.ts
@@ -10,6 +10,42 @@ const app = new Hono<{ Bindings: Env }>();
 
 app.get('/static/*', serveStatic({root: './', manifest}))
 app.get('/', c => c.html(layout(indexHtml())))
+
+const createEmbedding = async (ai: Ai, query: string): Promise<number[]> => {
+    const queryVector: EmbeddingResponse = await ai.run('@cf/baai/bge-large-en-v1.5', {
+        text: [query],
+    });
+    return queryVector.data[0];
+};
+
+const searchEmbeddings = async (index: VectorizeIndex, queryVector: number[]): Promise<string | undefined> => {
+    const result = await index.query(queryVector, {topK: 1, returnMetadata: true});
+    return result.matches[0]?.metadata?.link as string | undefined;
+};
+
+const retrieveArticle = async (store: KVNamespace, matchLink: string) => await store.get(matchLink);
+
+const fetchTextCompletion = async (ai: Ai, article: string, query: string): Promise<string> => {
+    const messages = [
+        {role: "system", content: "You are a reporter for a major world newspaper."},
+        {
+            role: "system",
+            content: `Use the following article as a source respond to the user's query: ${article.slice(0, 5500)}`
+        },
+        {
+            role: "system",
+            content: "Write your response as if you were writing a short, high-quality news article for your paper. Limit your response to one paragraph"
+        },
+        {
+            role: "user",
+            content: query,
+        },
+    ];
+
+    const textResult = await ai.run("@cf/meta/llama-3-8b-instruct", {messages});
+    return textResult.response;
+};
+
 app.post('/', async c => {
     const data = await c.req.formData()
     const query = data.get("query")
@@ -17,38 +53,21 @@ app.post('/', async c => {
         return c.html(layout(indexHtml({response: "Unable to answer query"})));
     }
 
-    const queryVector: EmbeddingResponse = await c.env.AI.run('@cf/baai/bge-large-en-v1.5', {
-        text: [query],
-    });
-
-    const result = await c.env.VECTORIZE_INDEX.query(queryVector.data[0], { topK: 1, returnMetadata: true });
-    const numberOfResults = result.matches.length;
-
-    const totalVectors = (await c.env.VECTORIZE_INDEX.describe()).vectorsCount
-    console.log(`found ${numberOfResults} results among ${totalVectors} vectors`)
+    const queryVector = await createEmbedding(c.env.AI, query);
 
-    const metadata = result.matches[0]?.metadata;
-    if (metadata === undefined) {
+    const source = await searchEmbeddings(c.env.VECTORIZE_INDEX, queryVector);
+    if (source === undefined) {
         return c.html(layout(indexHtml({query, response: "Unable to answer query"})));
     }
 
-    const matchLink = metadata.link as string;
-    const article = await c.env.ARTICLES.get(matchLink)
+    const article = await retrieveArticle(c.env.ARTICLES, source);
     if (article === null) {
-        return c.html(layout(indexHtml({query, response: "Unable to answer query", source: matchLink})));
+        return c.html(layout(indexHtml({query, response: "Unable to answer query", source: source})));
     }
-    const messages = [
-        { role: "system", content: "You are a reporter for a major world newspaper." },
-        { role: "system", content: `Use the following article as a source respond to the user's query: ${article.slice(0, 5500)}` },
-        { role: "system", content: "Write your response as if you were writing a short, high-quality news article for your paper. Limit your response to one paragraph" },
-        {
-            role: "user",
-            content: query,
-        },
-    ];
 
-    const textResult = await c.env.AI.run("@cf/meta/llama-3-8b-instruct", { messages });
-    return c.html(layout(indexHtml({query, response: textResult.response, source: matchLink})));
+    const response = await fetchTextCompletion(c.env.AI, article, query);
+
+    return c.html(layout(indexHtml({query, response, source})));
 })
 
 export default app satisfies ExportedHandler<Env>;