diff --git a/src/vs/platform/void/common/voidSettingsTypes.ts b/src/vs/platform/void/common/voidSettingsTypes.ts
index fce75bcd..64c74477 100644
--- a/src/vs/platform/void/common/voidSettingsTypes.ts
+++ b/src/vs/platform/void/common/voidSettingsTypes.ts
@@ -144,8 +144,8 @@ export const defaultProviderSettings = {
 export type ProviderName = keyof typeof defaultProviderSettings
 export const providerNames = Object.keys(defaultProviderSettings) as ProviderName[]
 
-export const localProviderNames: ProviderName[] = ['ollama'] // all local names
-export const nonlocalProviderNames = providerNames.filter((name) => !localProviderNames.includes(name)) // all non-local names
+export const localProviderNames = ['ollama', 'openAICompatible'] satisfies ProviderName[] // all local names
+export const nonlocalProviderNames = providerNames.filter((name) => !(localProviderNames as string[]).includes(name)) // all non-local names
 
 type CustomSettingName = UnionOfKeys<typeof defaultProviderSettings[ProviderName]>
 type CustomProviderSettings<providerName extends ProviderName> = {
@@ -379,7 +379,7 @@ export const featureNames = ['Ctrl+L', 'Ctrl+K', 'Autocomplete'] as const
 
 
 // the models of these can be refreshed (in theory all can, but not all should)
-export const refreshableProviderNames = ['ollama', 'openAICompatible'] satisfies ProviderName[]
+export const refreshableProviderNames = localProviderNames
 export type RefreshableProviderName = typeof refreshableProviderNames[number]
 
 
@@ -405,7 +405,7 @@ type FeatureFlagDisplayInfo = {
 export const displayInfoOfFeatureFlag = (featureFlag: FeatureFlagName): FeatureFlagDisplayInfo => {
 	if (featureFlag === 'autoRefreshModels') {
 		return {
-			description: `Automatically detect local providers and models (like Ollama).`, // ${`refreshableProviderNames.map(providerName => titleOfProviderName(providerName)).join(', ')`}
+			description: `Automatically detect local providers and models (${refreshableProviderNames.map(providerName => displayInfoOfProviderName(providerName).title).join(', ')}).`,
 		}
 	}
 	throw new Error(`featureFlagInfo: Unknown feature flag: "${featureFlag}"`)
diff --git a/src/vs/workbench/contrib/void/browser/inlineDiffsService.ts b/src/vs/workbench/contrib/void/browser/inlineDiffsService.ts
index da288609..a1c7fd65 100644
--- a/src/vs/workbench/contrib/void/browser/inlineDiffsService.ts
+++ b/src/vs/workbench/contrib/void/browser/inlineDiffsService.ts
@@ -117,7 +117,7 @@ type CtrlKZone = {
 	editorId: string; // the editor the input lives on
 
 	_mountInfo: null | {
-		inputBox: InputBox | null; // the input box that lives in the zone
+		inputBoxRef: { current: InputBox | null }; // the input box that lives in the zone
 		dispose: () => void;
 		refresh: () => void;
 	}
@@ -329,7 +329,7 @@ class InlineDiffsService extends Disposable implements IInlineDiffsService {
 
 		let zoneId: string | null = null
 		let viewZone_: IViewZone | null = null
-		let inputBox_: InputBox | null = null
+		const inputBoxRef: { current: InputBox | null } = { current: null }
 
 		const itemId = this._consistentEditorItemService.addToEditor(editor, () => {
 			const domNode = document.createElement('div');
@@ -352,7 +352,7 @@ class InlineDiffsService extends Disposable implements IInlineDiffsService {
 				mountCtrlK(domNode, accessor, {
 					diffareaid: ctrlKZone.diffareaid,
 					onGetInputBox: (inputBox) => {
-						inputBox_ = inputBox
+						inputBoxRef.current = inputBox
 						// if it's mounting for the first time, focus it
 						if (!(ctrlKZone.diffareaid in this.mostRecentTextOfCtrlKZoneId)) { // detect first mount this way (a hack)
 							this.mostRecentTextOfCtrlKZoneId[ctrlKZone.diffareaid] = undefined
@@ -381,10 +381,8 @@ class InlineDiffsService extends Disposable implements IInlineDiffsService {
 			})
 		})
 
-
-
 		return {
-			inputBox: inputBox_,
+			inputBoxRef,
 			refresh: () => editor.changeViewZones(accessor => {
 				if (zoneId && viewZone_) {
 					viewZone_.afterLineNumber = ctrlKZone.startLine - 1
@@ -394,7 +392,7 @@ class InlineDiffsService extends Disposable implements IInlineDiffsService {
 			dispose: () => {
 				this._consistentEditorItemService.removeFromEditor(itemId)
 			},
-		}
+		} satisfies CtrlKZone['_mountInfo']
 	}
 
 
@@ -405,6 +403,7 @@ class InlineDiffsService extends Disposable implements IInlineDiffsService {
 			if (diffArea.type !== 'CtrlKZone') continue
 			if (!diffArea._mountInfo) {
 				diffArea._mountInfo = this._addCtrlKZoneInput(diffArea)
+				console.log('MOUNTED', diffArea.diffareaid)
 			}
 			else {
 				diffArea._mountInfo.refresh()
@@ -887,7 +886,7 @@ class InlineDiffsService extends Disposable implements IInlineDiffsService {
 			if (diffArea.type !== 'CtrlKZone') continue
 			const noOverlap = diffArea.startLine > endLine || diffArea.endLine < startLine
 			if (!noOverlap) {
-				setTimeout(() => diffArea._mountInfo?.inputBox?.focus(), 0)
+				setTimeout(() => diffArea._mountInfo?.inputBoxRef.current?.focus(), 0)
 				return
 			}
 		}
@@ -943,6 +942,8 @@ class InlineDiffsService extends Disposable implements IInlineDiffsService {
 		let uri: URI
 		let userMessage: string
 
+		console.log('AA')
+
 		if (featureName === 'Ctrl+L') {
 
 			const uri_ = this._getActiveEditorURI()
@@ -983,14 +984,13 @@ class InlineDiffsService extends Disposable implements IInlineDiffsService {
 			startLine = startLine_
 			endLine = endLine_
 
-			if (!_mountInfo?.inputBox) return
-			userMessage = _mountInfo.inputBox?.value
+			if (!_mountInfo?.inputBoxRef.current) return
+			userMessage = _mountInfo.inputBoxRef.current?.value
 		}
 		else {
 			throw new Error(`Void: diff.type not recognized on: ${featureName}`)
 		}
 
-
 		const currentFileStr = this._readURI(uri)
 		if (currentFileStr === null) return
 		const originalCode = currentFileStr.split('\n').slice((startLine - 1), (endLine - 1) + 1).join('\n')
diff --git a/src/vs/workbench/contrib/void/browser/react/src/void-settings-tsx/Settings.tsx b/src/vs/workbench/contrib/void/browser/react/src/void-settings-tsx/Settings.tsx
index 62fd2450..cc935cc3 100644
--- a/src/vs/workbench/contrib/void/browser/react/src/void-settings-tsx/Settings.tsx
+++ b/src/vs/workbench/contrib/void/browser/react/src/void-settings-tsx/Settings.tsx
@@ -420,13 +420,12 @@ export const Settings = () => {
 							<h2 className={`text-3xl mb-2`}>Local Providers</h2>
 							{/* <h3 className={`text-md opacity-50 mb-2`}>{`Keep your data private by hosting AI locally on your computer.`}</h3> */}
 							{/* <h3 className={`text-md opacity-50 mb-2`}>{`Instructions:`}</h3> */}
-							<h3 className={`text-md opacity-50 mb-2`}>{`Void can access any model that you host locally.`}</h3>
-							<div className='pl-4 select-text'>
-								<h4 className={`text-xs opacity-50 mb-2`}><ChatMarkdownRender string={`1. Download [Ollama](https://ollama.com/download).`} /></h4>
-								<h4 className={`text-xs opacity-50 mb-2`}><ChatMarkdownRender string={`2. Open your terminal.`} /></h4>
-								<h4 className={`text-xs opacity-50 mb-2`}><ChatMarkdownRender string={`3. Run \`ollama run llama3.1\`. This installs Meta's llama model which is competitive with GPT-series models, and requires 5GB of memory.`} /></h4>
-								<h4 className={`text-xs opacity-50 mb-2`}><ChatMarkdownRender string={`4. Run \`ollama run qwen2.5-coder:1.5b\`. This is a faster autocomplete model and requires 1GB of memory.`} /></h4>
-								<h4 className={`text-xs opacity-50 mb-2`}><ChatMarkdownRender string={`5. Void will automatically detect your Ollama models. You can customize the endpoint and models below.`} /></h4>
+							<h3 className={`text-md mb-2`}>{`Void can access any model that you host locally. By default, we automatically detect your local models.`}</h3>
+							<div className='pl-4 select-text opacity-50'>
+								<h4 className={`text-xs mb-2`}><ChatMarkdownRender string={`1. Download [Ollama](https://ollama.com/download).`} /></h4>
+								<h4 className={`text-xs mb-2`}><ChatMarkdownRender string={`2. Open your terminal.`} /></h4>
+								<h4 className={`text-xs mb-2`}><ChatMarkdownRender string={`3. Run \`ollama run llama3.1\`. This installs Meta's llama model which is competitive with GPT-series models. It requires 5GB of memory.`} /></h4>
+								<h4 className={`text-xs mb-2`}><ChatMarkdownRender string={`4. Run \`ollama run qwen2.5-coder:1.5b\`. This is a faster autocomplete model and requires 1GB of memory.`} /></h4>
 								{/* TODO we should create UI for downloading models without user going into terminal */}
 							</div>
 
@@ -435,7 +434,7 @@ export const Settings = () => {
 							</ErrorBoundary>
 
 							<h2 className={`text-3xl mb-2 mt-16`}>More Providers</h2>
-							<h3 className={`text-md opacity-50 mb-2`}>{`Void can also access models like ChatGPT and Claude. We recommend using Anthropic or OpenAI.`}</h3>
+							<h3 className={`text-md mb-2`}>{`Void can also access models like ChatGPT and Claude. We recommend using Anthropic or OpenAI.`}</h3>
 							{/* <h3 className={`text-md opacity-50 mb-2`}>{`Access models like ChatGPT and Claude. We recommend using Anthropic or OpenAI as providers, or Groq as a faster alternative.`}</h3> */}
 							<ErrorBoundary>
 								<VoidProviderSettings providerNames={nonlocalProviderNames} />