diff --git a/.changeset/social-mammals-chew.md b/.changeset/social-mammals-chew.md
new file mode 100644
index 00000000..e6f35b71
--- /dev/null
+++ b/.changeset/social-mammals-chew.md
@@ -0,0 +1,7 @@
+---
+'workers-observability': minor
+'workers-bindings': minor
+'docs-vectorize': minor
+---
+
+Updated the model for docs search to embeddinggemma-300m
diff --git a/apps/docs-vectorize/wrangler.jsonc b/apps/docs-vectorize/wrangler.jsonc
index bf99ac86..90c8edf1 100644
--- a/apps/docs-vectorize/wrangler.jsonc
+++ b/apps/docs-vectorize/wrangler.jsonc
@@ -31,7 +31,7 @@
 	"vectorize": [
 		{
 			"binding": "VECTORIZE",
-			"index_name": "docs-bge-base"
+			"index_name": "docs-embeddinggemma-v1"
 		}
 	],
 	"vars": {
@@ -72,7 +72,7 @@
 			"vectorize": [
 				{
 					"binding": "VECTORIZE",
-					"index_name": "docs-bge-base"
+					"index_name": "docs-embeddinggemma-v1"
 				}
 			],
 			"analytics_engine_datasets": [
@@ -103,7 +103,7 @@
 			"vectorize": [
 				{
 					"binding": "VECTORIZE",
-					"index_name": "docs-bge-base"
+					"index_name": "docs-embeddinggemma-v1"
 				}
 			],
 			"analytics_engine_datasets": [
diff --git a/apps/workers-bindings/wrangler.jsonc b/apps/workers-bindings/wrangler.jsonc
index e171a91e..1e71e5dc 100644
--- a/apps/workers-bindings/wrangler.jsonc
+++ b/apps/workers-bindings/wrangler.jsonc
@@ -51,7 +51,7 @@
 	"vectorize": [
 		{
 			"binding": "VECTORIZE",
-			"index_name": "docs-bge-base"
+			"index_name": "docs-embeddinggemma-v1"
 		}
 	],
 	"workers_dev": false,
@@ -92,7 +92,7 @@
 			"vectorize": [
 				{
 					"binding": "VECTORIZE",
-					"index_name": "docs-bge-base"
+					"index_name": "docs-embeddinggemma-v1"
 				}
 			],
 			"vars": {
@@ -141,7 +141,7 @@
 			"vectorize": [
 				{
 					"binding": "VECTORIZE",
-					"index_name": "docs-bge-base"
+					"index_name": "docs-embeddinggemma-v1"
 				}
 			],
 			"analytics_engine_datasets": [
diff --git a/apps/workers-observability/wrangler.jsonc b/apps/workers-observability/wrangler.jsonc
index 2eaa2c95..8d60496e 100644
--- a/apps/workers-observability/wrangler.jsonc
+++ b/apps/workers-observability/wrangler.jsonc
@@ -41,7 +41,7 @@
 	"vectorize": [
 		{
 			"binding": "VECTORIZE",
-			"index_name": "docs-bge-base"
+			"index_name": "docs-embeddinggemma-v1"
 		}
 	],
 	"vars": {
@@ -98,7 +98,7 @@
 			"vectorize": [
 				{
 					"binding": "VECTORIZE",
-					"index_name": "docs-bge-base"
+					"index_name": "docs-embeddinggemma-v1"
 				}
 			],
 			"analytics_engine_datasets": [
@@ -136,7 +136,7 @@
 			"vectorize": [
 				{
 					"binding": "VECTORIZE",
-					"index_name": "docs-bge-base"
+					"index_name": "docs-embeddinggemma-v1"
 				}
 			],
 			"vars": {
diff --git a/packages/mcp-common/src/tools/docs-vectorize.tools.ts b/packages/mcp-common/src/tools/docs-vectorize.tools.ts
index 0079f7ad..0395a777 100644
--- a/packages/mcp-common/src/tools/docs-vectorize.tools.ts
+++ b/packages/mcp-common/src/tools/docs-vectorize.tools.ts
@@ -42,6 +42,7 @@ export function registerDocsTools(agent: CloudflareMcpAgentNoAccount, env: Requi
 				.map((result) => {
 					return `<result>
 <url>${result.url}</url>
+<title>${result.title}</title>
 <text>
 ${result.text}
 </text>
@@ -93,10 +94,8 @@ ${result.text}
 }
 
 async function queryVectorize(ai: Ai, vectorizeIndex: VectorizeIndex, query: string, topK: number) {
-	// Recommendation from: https://huggingface.co/BAAI/bge-base-en-v1.5#model-list
-	const [queryEmbedding] = await getEmbeddings(ai, [
-		'Represent this sentence for searching relevant passages: ' + query,
-	])
+	// Recommendation from: https://ai.google.dev/gemma/docs/embeddinggemma/model_card#prompt_instructions
+	const [queryEmbedding] = await getEmbeddings(ai, ['task: search result | query: ' + query])
 
 	const { matches } = await vectorizeIndex.query(queryEmbedding, {
 		topK,
@@ -108,6 +107,7 @@ async function queryVectorize(ai: Ai, vectorizeIndex: VectorizeIndex, query: str
 		similarity: Math.min(match.score, 1),
 		id: match.id,
 		url: sourceToUrl(String(match.metadata?.filePath ?? '')),
+		title: String(match.metadata?.title ?? ''),
 		text: String(match.metadata?.text ?? ''),
 	}))
 }
@@ -123,15 +123,15 @@ function sourceToUrl(path: string) {
 	)
 }
 
-async function getEmbeddings(ai: Ai, strings: string[]) {
+async function getEmbeddings(ai: Ai, strings: string[]): Promise<number[][]> {
 	const response = await doWithRetries(() =>
-		ai.run('@cf/baai/bge-base-en-v1.5', {
+		// @ts-expect-error embeddinggemma not in types yet
+		ai.run('@cf/google/embeddinggemma-300m', {
 			text: strings,
-			// @ts-expect-error pooling not in types yet
-			pooling: 'cls',
 		})
 	)
 
+	// @ts-expect-error embeddinggemma not in types yet
 	return response.data
 }