fix: use fallback model for name generation when preferred models unavailable

If Haiku/GPT-Mini aren't available (no Anthropic/OpenAI key), fall back
to the user's configured message model. This avoids confusing errors when
users have only configured alternative providers.

Author: ammar-agent

src/browser/components/ChatInput/useCreationWorkspace.ts

@@ -102,18 +102,22 @@ export function useCreationWorkspace({
     getRuntimeString,
   } = useDraftWorkspaceSettings(projectPath, branches, recommendedTrunk);
 
+  // Project scope ID for reading send options at send time
+  const projectScopeId = getProjectScopeId(projectPath);
+
+  // Read the user's preferred model for fallback (same source as ChatInput's preferredModel)
+  const fallbackModel = readPersistedState<string | null>(getModelKey(projectScopeId), null);
+
   // Workspace name generation with debounce
   const workspaceNameState = useWorkspaceName({
     message,
     debounceMs: 500,
+    fallbackModel: fallbackModel ?? undefined,
   });
 
   // Destructure name state functions for use in callbacks
   const { waitForGeneration } = workspaceNameState;
 
-  // Project scope ID for reading send options at send time
-  const projectScopeId = getProjectScopeId(projectPath);
-
   // Load branches on mount
   useEffect(() => {
     // This can be created with an empty project path when the user is

src/browser/hooks/useWorkspaceName.ts

@@ -6,6 +6,8 @@ export interface UseWorkspaceNameOptions {
   message: string;
   /** Debounce delay in milliseconds (default: 500) */
   debounceMs?: number;
+  /** Model to use if preferred small models aren't available */
+  fallbackModel?: string;
 }
 
 /** State and actions for workspace name generation, suitable for passing to components */
@@ -37,7 +39,7 @@ export interface UseWorkspaceNameReturn extends WorkspaceNameState {
  * auto-generation resumes.
  */
 export function useWorkspaceName(options: UseWorkspaceNameOptions): UseWorkspaceNameReturn {
-  const { message, debounceMs = 500 } = options;
+  const { message, debounceMs = 500, fallbackModel } = options;
   const { api } = useAPI();
 
   const [generatedName, setGeneratedName] = useState("");
@@ -103,6 +105,7 @@ export function useWorkspaceName(options: UseWorkspaceNameOptions): UseWorkspace
       try {
         const result = await api.nameGeneration.generate({
           message: forMessage,
+          fallbackModel,
         });
 
         // Check if this request is still current (wasn't cancelled)
@@ -141,7 +144,7 @@ export function useWorkspaceName(options: UseWorkspaceNameOptions): UseWorkspace
         }
       }
     },
-    [api]
+    [api, fallbackModel]
   );
 
   // Debounced generation effect

src/common/orpc/schemas/api.ts

@@ -278,6 +278,8 @@ export const nameGeneration = {
   generate: {
     input: z.object({
       message: z.string(),
+      /** Model to use if preferred small models (Haiku, GPT-Mini) aren't available */
+      fallbackModel: z.string().optional(),
     }),
     output: ResultSchema(
       z.object({

src/node/orpc/router.ts

@@ -174,13 +174,15 @@ export const router = (authToken?: string) => {
         .input(schemas.nameGeneration.generate.input)
         .output(schemas.nameGeneration.generate.output)
         .handler(async ({ context, input }) => {
-          const model = await getPreferredNameModel(context.aiService);
+          // Prefer small/fast models, fall back to user's configured model
+          const model =
+            (await getPreferredNameModel(context.aiService)) ?? input.fallbackModel;
           if (!model) {
             return {
               success: false,
               error: {
                 type: "unknown" as const,
-                raw: "No model available for name generation. Configure an API key for Anthropic or OpenAI.",
+                raw: "No model available for name generation.",
               },
             };
           }