Update rag.mdx

docs/troubleshooting/rag.mdx

@@ -3,15 +3,15 @@ sidebar_position: 3
 title: "🧠 Troubleshooting RAG (Retrieval-Augmented Generation)"
 ---
 
-Retrieval-Augmented Generation (RAG) enables language models to reason over external content—documents, knowledge bases, and more—by retrieving relevant info and feeding it into the model. But when things don’t work as expected (e.g., the model "hallucinates" or misses relevant info), it's often not the model's fault—it's a context issue.
+Retrieval-Augmented Generation (RAG) enables language models to reason over external content—documents, knowledge bases, and more—by retrieving relevant info and feeding it into the model. But when things don't work as expected (e.g., the model "hallucinates" or misses relevant info), it's often not the model's fault—it's a context issue.
 
-Let’s break down the common causes and solutions so you can supercharge your RAG accuracy! 🚀
+Let's break down the common causes and solutions so you can supercharge your RAG accuracy! 🚀
 
 ## Common RAG Issues and How to Fix Them 🛠️
 
-### 1. The Model "Can’t See" Your Content 👁️❌
+### 1. The Model "Can't See" Your Content 👁️❌
 
-This is the most common problem—and it's typically caused by issues during your content ingestion process. The model doesn’t hallucinate because it’s wrong, it hallucinates because it was never given the right content in the first place.
+This is the most common problem—and it's typically caused by issues during your content ingestion process. The model doesn't hallucinate because it's wrong, it hallucinates because it was never given the right content in the first place.
 
 ✅ Solution: Check your content extraction settings
 
@@ -21,7 +21,7 @@ This is the most common problem—and it's typically caused by issues during you
   - Docling
   - Custom extractors (depending on your document types)
 
-📌 Tip: Try uploading a document and preview the extracted content. If it’s blank or missing key sections, you need to adjust your extractor settings or use a different engine.
+📌 Tip: Try uploading a document and preview the extracted content. If it's blank or missing key sections, you need to adjust your extractor settings or use a different engine.
 
 ---
 
@@ -33,33 +33,50 @@ Open WebUI is designed to work with models that have limited context windows by
 
 - Go to **Admin Settings > Documents**
 - Either:
-  - 💡 Enable “Bypass Embedding and Retrieval” — This sends full content directly without applying strict retrieval filters.
-  - 🔍 Toggle on “Full Context Mode” — This injects more comprehensive content into the model prompt.
+  - 💡 Enable "Bypass Embedding and Retrieval" — This sends full content directly without applying strict retrieval filters.
+  - 🔍 Toggle on "Full Context Mode" — This injects more comprehensive content into the model prompt.
 
-📌 Warning: Be mindful of context limits—if your model can’t handle more tokens, it will still get cut off.
+📌 Warning: Be mindful of context limits—if your model can't handle more tokens, it will still get cut off.
 
 ---
 
 ### 3. Token Limit is Too Short ⏳
 
-Even if retrieval works, your model might still not process all the content it receives—because it simply can’t.
+Even if retrieval works, your model might still not process all the content it receives—because it simply can't.
 
 By default, many models (especially Ollama-hosted LLMs) are limited to a 2048-token context window. That means only a fraction of your retrieved data will actually be used.
 
+**Why Web Search Especially Needs Larger Context Windows:**
+Web pages are particularly challenging for small context windows because they contain far more content than typical documents. A single web page often includes:
+- Main content (the actual information you want)
+- Navigation menus, headers, and footers
+- Sidebar content and advertisements
+- Comments sections and related links
+- Metadata and embedded scripts
+
+Even after content extraction and cleaning, web pages easily consume 4,000-8,000+ tokens of context. With a 2048-token limit, you're getting less than half the content, often missing the most relevant information that appears later in the page. Even 4096 tokens is frequently insufficient for comprehensive web content analysis.
+
 ✅ Solutions:
 
-- 🛠️ Extend the model’s context length:
-  - Navigate to the **Model Editor or Chat Controls**
-  - Modify the context length (e.g., increase to 8192+ tokens if supported)
+- 🛠️ **For Ollama Models**: Extend the model's context length:
+  - Navigate to: **Admin Panel > Models > Settings** (of the model you want to edit)
+  - Go to **Advanced Parameters**
+  - Modify the context length (e.g., increase to 8192+ tokens if supported by your model)
 
-ℹ️ Note: The 2048-token default is a big limiter. For better RAG results, we recommend using models that support longer contexts.
+- 🌐 **For OpenAI and Other Integrated Models**: These models typically have their own context limits that cannot be modified through Open WebUI settings. Ensure you're using a model with sufficient context length:
+  - GPT-4: 8,192 tokens
+  - GPT-4-32k: 32,768 tokens
+  - GPT-4 Turbo: 128,000 tokens
+  - Claude 3: Up to 200,000 tokens
+
+ℹ️ Note: The 2048-token default is a big limiter for web search. For better RAG results with web content, we strongly recommend using at least 8192 tokens, with 16384+ being ideal for complex web pages.
 
 ✅ Alternative: Use an external LLM with larger context capacity
 
 - Try GPT-4, GPT-4o, Claude 3, Gemini 1.5, or Mixtral with 8k+ context
-- Compare performance to Ollama—notice the accuracy difference when more content can be injected!
+- Compare performance to Ollama—notice the dramatic accuracy difference when more web content can be processed!
 
-📌 Tip: Stick with external models for better RAG performance in production use cases.
+📌 Tip: For web search and complex document analysis, stick with models that support 8192+ token contexts in production use cases.
 
 ---
 
@@ -80,10 +97,10 @@ Bad embeddings = bad retrieval. If the vector representation of your content is
 
 ### 5. ❌ 400: 'NoneType' object has no attribute 'encode'
 
-This error indicates a misconfigured or missing embedding model. When Open WebUI tries to create embeddings but doesn’t have a valid model loaded, it can’t process the text—and the result is this cryptic error.
+This error indicates a misconfigured or missing embedding model. When Open WebUI tries to create embeddings but doesn't have a valid model loaded, it can't process the text—and the result is this cryptic error.
 
 💥 Cause:
-- Your embedding model isn’t set up properly.
+- Your embedding model isn't set up properly.
 - It might not have downloaded completely.
 - Or if you're using an external embedding model, it may not be accessible.
 
@@ -99,7 +116,7 @@ This error indicates a misconfigured or missing embedding model. When Open WebUI
 
 ## 🧪 Pro Tip: Test with GPT-4o or GPT-4
 
-If you’re not sure whether the issue is with retrieval, token limits, or embedding—try using GPT-4o temporarily (e.g., via OpenAI API). If the results suddenly become more accurate, it's a strong signal that your local model’s context limit (2048 by default in Ollama) is the bottleneck.
+If you're not sure whether the issue is with retrieval, token limits, or embedding—try using GPT-4o temporarily (e.g., via OpenAI API). If the results suddenly become more accurate, it's a strong signal that your local model's context limit (2048 by default in Ollama) is the bottleneck.
 
 - GPT-4o handles larger inputs (128k tokens!)
 - Provides a great benchmark to evaluate your system's RAG reliability
@@ -110,12 +127,12 @@ If you’re not sure whether the issue is with retrieval, token limits, or embed
 
 | Problem | Fix |
 |--------|------|
-| 🤔 Model can’t “see” content | Check document extractor settings |
+| 🤔 Model can't "see" content | Check document extractor settings |
 | 🧹 Only part of content used | Enable Full Context Mode or Bypass Embedding |
-| ⏱ Limited by 2048 token cap | Increase model context length or use large-context LLM |
+| ⏱ Limited by 2048 token cap | Increase model context length (Admin Panel > Models > Settings > Advanced Parameters for Ollama) or use large-context LLM |
 | 📉 Inaccurate retrieval | Switch to a better embedding model, then reindex |
 | Still confused? | Test with GPT-4o and compare outputs |
 
 ---
 
-By optimizing these areas—extraction, embedding, retrieval, and model context—you can dramatically improve how accurately your LLM works with your documents. Don’t let a 2048-token window or weak retrieval pipeline hold back your AI’s power 🎯.
+By optimizing these areas—extraction, embedding, retrieval, and model context—you can dramatically improve how accurately your LLM works with your documents. Don't let a 2048-token window or weak retrieval pipeline hold back your AI's power 🎯.