feat: Removes 'excludeThoughtOnReq' setting

Removes the 'excludeThoughtOnReq' setting, which was intended to exclude the thought process from API requests.

The reasoning format is now always requested from the server but is not sent back in API requests.

Author: allozaur

tools/server/webui/src/lib/components/app/chat/ChatSettings/ChatSettingsDialog.svelte

@@ -202,11 +202,6 @@
 					key: 'showThoughtInProgress',
 					label: 'Show thought in progress',
 					type: 'checkbox'
-				},
-				{
-					key: 'excludeThoughtOnReq',
-					label: 'Exclude thought on request',
-					type: 'checkbox'
 				}
 			]
 		},

tools/server/webui/src/lib/constants/settings-config.ts

@@ -6,7 +6,6 @@ export const SETTING_CONFIG_DEFAULT: Record<string, string | number | boolean> =
 	theme: 'system',
 	showTokensPerSecond: false,
 	showThoughtInProgress: true,
-	excludeThoughtOnReq: false,
 	pasteLongTextToFileLen: 2500,
 	pdfAsImage: false,
 	// make sure these default values are in sync with `common.h`
@@ -61,7 +60,6 @@ export const SETTING_CONFIG_INFO: Record<string, string> = {
 	custom: 'Custom JSON parameters to send to the API. Must be valid JSON format.',
 	showTokensPerSecond: 'Display generation speed in tokens per second during streaming.',
 	showThoughtInProgress: 'Expand thought process by default when generating messages.',
-	excludeThoughtOnReq: 'Exclude thought process when sending requests to API (Recommended for DeepSeek-R1).',
 	pdfAsImage: 'Parse PDF as image instead of text (requires vision-capable model).',
 	pyInterpreterEnabled: 'Enable Python interpreter using Pyodide. Allows running Python code in markdown code blocks.'
 };

tools/server/webui/src/lib/services/chat.ts

@@ -40,7 +40,6 @@ export class ChatService {
 
 		// Build base request body with system message injection
 		const processedMessages = this.injectSystemMessage(messages);
-		const currentConfig = config();
 
 		const requestBody: ApiChatCompletionRequest = {
 			messages: processedMessages.map((msg: ApiChatMessageData) => ({
@@ -50,9 +49,8 @@ export class ChatService {
 			stream
 		};
 
-		if (!currentConfig.excludeThoughtOnReq) {
-			requestBody.reasoning_format = 'auto';
-		}
+		// Always request reasoning content from server, but never send it back in message history
+		requestBody.reasoning_format = 'auto';
 
 		// Add generation parameters if provided
 		if (temperature !== undefined) requestBody.temperature = temperature;
@@ -195,7 +193,6 @@ export class ChatService {
 		const decoder = new TextDecoder();
 		let fullResponse = '';
 		let fullReasoningContent = '';
-		let thinkContent = '';
 		let regularContent = '';
 		let insideThinkTag = false;
 		let hasReceivedData = false;
@@ -240,8 +237,8 @@ export class ChatService {
 								insideThinkTag = this.processContentForThinkTags(
 									content,
 									insideThinkTag,
-									(thinkChunk) => {
-										thinkContent += thinkChunk;
+									() => {
+										// Think content is ignored - we don't include it in API requests
 									},
 									(regularChunk) => {
 										regularContent += regularChunk;