From 27b0dbabbb05aa145aa64f1f950cd7e95079b2fd Mon Sep 17 00:00:00 2001
From: Anush <mail@anush.sh>
Date: Tue, 5 May 2026 11:29:49 +0530
Subject: [PATCH] docs: Updated gemini.md for multiple inputs (#2330)

---
 .../documentation/embeddings/gemini.md        | 34 ++++++++++++-------
 1 file changed, 22 insertions(+), 12 deletions(-)

diff --git a/qdrant-landing/content/documentation/embeddings/gemini.md b/qdrant-landing/content/documentation/embeddings/gemini.md
index d1fe19c90..87dca23a4 100644
--- a/qdrant-landing/content/documentation/embeddings/gemini.md
+++ b/qdrant-landing/content/documentation/embeddings/gemini.md
@@ -39,11 +39,14 @@ texts = [
     "Gemini is a family of natively multimodal, large language models (LLMs).",
 ]
 
-result = gemini_client.models.embed_content(
-    model="gemini-embedding-2",
-    contents=texts,
-    config=types.EmbedContentConfig(task_type="RETRIEVAL_DOCUMENT"),
-)
+embeddings = [
+    gemini_client.models.embed_content(
+        model="gemini-embedding-2",
+        contents=text,
+        config=types.EmbedContentConfig(task_type="RETRIEVAL_DOCUMENT"),
+    ).embeddings[0]
+    for text in texts
+]
 ```
 
 ```typescript
@@ -58,13 +61,20 @@ const texts = [
   "Gemini is a family of natively multimodal, large language models (LLMs).",
 ];
 
-const result = await geminiClient.models.embedContent({
-  model: "gemini-embedding-2",
-  contents: texts,
-  config: { taskType: "RETRIEVAL_DOCUMENT" },
-});
+const embeddings = await Promise.all(
+  texts.map(async (text) => {
+    const result = await geminiClient.models.embedContent({
+      model: "gemini-embedding-2",
+      contents: text,
+      config: { taskType: "RETRIEVAL_DOCUMENT" },
+    });
+    return result.embeddings[0];
+  })
+);
 ```
 
+> Note: `gemini-embedding-2` returns one aggregated embedding when given multiple inputs. Embed each text separately and parallelize on the caller side.
+
 ## Creating Qdrant Points and Indexing documents with Qdrant
 
 ### Creating Qdrant Points
@@ -76,14 +86,14 @@ points = [
         vector=embedding.values,
         payload={"text": text},
     )
-    for idx, (embedding, text) in enumerate(zip(result.embeddings, texts))
+    for idx, (embedding, text) in enumerate(zip(embeddings, texts))
 ]
 ```
 
 ```typescript
 const points = texts.map((text, idx) => ({
   id: idx,
-  vector: result.embeddings[idx].values,
+  vector: embeddings[idx].values,
   payload: { text },
 }));
 ```