better doc

.gitignore

@@ -1,6 +1,7 @@
 # Dependencies
 node_modules/
 /package-lock.json
+.pnpm-store/
 
 # Turborepo
 .turbo

docs/index.html

@@ -407,10 +407,10 @@ <h2>Documentation</h2>
                 <p>Complete API reference and guides</p>
             </div>
             <div class="docs-grid">
-                <a href="llms-full.md" class="doc-card doc-card-llm">
+                <a href="llms-full.txt" class="doc-card doc-card-llm">
                     <div class="doc-icon">🤖</div>
                     <h3>Full LLM friendly Documentation</h3>
-                    <p>Comprehensive llms-full.md for LLMs, AI agents, and vibe coding — copy the entire file into your
+                    <p>Comprehensive llms-full.txt for LLMs, AI agents, and vibe coding — copy the entire file into your
                         context</p>
                 </a>
                 <a href="https://github.com/dh7/mindcache/blob/main/packages/mindcache/src/react/README.md" class="doc-card">

docs/llms-full.md → docs/llms-full.txt

@@ -1,4 +1,3 @@
-{% raw %}
 # MindCache - Complete LLM Documentation
 
 > This document is optimized for LLMs to understand and work with MindCache. It contains comprehensive API documentation with extensive code examples.
@@ -2480,5 +2479,3 @@ NEXT_PUBLIC_INSTANCE_ID=your-instance-id
 ---
 
 *This documentation is optimized for LLM consumption. For human-readable documentation with interactive examples, visit [mindcache.dev](https://mindcache.dev).*
-{% endraw %}
-

docs/llms.txt

@@ -7,7 +7,7 @@ MindCache 3.7 provides persistent memory for AI agents with drop-in React compon
 ## Quick Links
 
 - Documentation: https://mindcache.dev
-- Full LLM Documentation: https://mindcache.dev/llms-full.md
+- Full LLM Documentation: https://mindcache.dev/llms-full.txt
 - GitHub: https://github.com/dh7/mindcache
 - npm: https://www.npmjs.com/package/mindcache
 
@@ -139,5 +139,5 @@ const { status, save, load } = useLocalFirstSync({ mindcache, gitstore: { owner,
 
 ## Full Documentation
 
-See https://mindcache.dev/llms-full.md
+See https://mindcache.dev/llms-full.txt
 

packages/mindcache/README.md

@@ -5,7 +5,7 @@ A TypeScript library for managing short-term memory in AI agents through an LLM-
 ## Documentation
 
 *   **[API Reference](./docs/mindcache-api.md)**: Detailed method signatures and type definitions. Optimized for AI agents (like Cursor, Claude, etc.) to understand the library's capabilities.
-*   **[Full Documentation](https://mindcache.dev/llms-full.md)**: Comprehensive guide with examples and deep dives.
+*   **[Full Documentation](https://mindcache.dev/llms-full.txt)**: Comprehensive guide with examples and deep dives.
 
 ## Quick Start
 

packages/mindcache/package.json

@@ -1,6 +1,6 @@
 {
   "name": "mindcache",
-  "version": "3.8.0",
+  "version": "3.8.1",
   "description": "A TypeScript library for managing short-term memory in AI agents through an LLM-friendly key-value repository",
   "main": "./dist/index.js",
   "module": "./dist/index.mjs",

packages/mindcache/src/core/MindCache.ts

@@ -681,7 +681,8 @@ export class MindCache {
       if (!this._idbProvider) {
         return resolve();
       }
-      this._idbProvider.on('synced', () => {
+      // Cast needed: IndexeddbPersistence extends Observable which has 'on' method
+      (this._idbProvider as unknown as { on: (event: string, cb: () => void) => void }).on('synced', () => {
         this._isLoaded = true;
         resolve();
       });

packages/mindcache/src/react/useClientChat.ts

@@ -189,6 +189,9 @@ export function useClientChat(options: UseClientChatOptions = {}): UseClientChat
     setError(null);
     setStreamingContent('');
 
+    // Track accumulated text for abort handling
+    let accumulatedText = '';
+
     try {
       // Get model from context (handles provider config automatically)
       const model = context.getModel();
@@ -207,7 +210,6 @@ export function useClientChat(options: UseClientChatOptions = {}): UseClientChat
 
       // Accumulated parts for the final message
       const parts: MessagePart[] = [];
-      let accumulatedText = '';
 
       // Stream the response with real-time updates
       const result = await streamText({
@@ -250,11 +252,7 @@ export function useClientChat(options: UseClientChatOptions = {}): UseClientChat
               }
             }
           }
-
-          // Add text from this step if any
-          if (step.text) {
-            accumulatedText += step.text;
-          }
+          // Note: Don't accumulate step.text here - it's already in textStream
         }
       });
 
@@ -264,16 +262,19 @@ export function useClientChat(options: UseClientChatOptions = {}): UseClientChat
         setStreamingContent(accumulatedText);
       }
 
+      // Get final text from result (authoritative, avoids any streaming duplication)
+      const finalText = await result.text;
+
       // Build final message with parts
-      if (accumulatedText) {
-        parts.unshift({ type: 'text', text: accumulatedText });
+      if (finalText) {
+        parts.unshift({ type: 'text', text: finalText });
       }
 
       // Add assistant message with all parts
       const assistantMessage: ChatMessage = {
         id: generateId(),
         role: 'assistant',
-        content: accumulatedText,
+        content: finalText,
         parts: parts.length > 0 ? parts : undefined,
         createdAt: new Date()
       };
@@ -284,13 +285,18 @@ export function useClientChat(options: UseClientChatOptions = {}): UseClientChat
       onFinish?.(assistantMessage);
 
     } catch (err) {
-      if ((err as Error).name === 'AbortError') {
-        // If we have partial content, save it
-        if (streamingContent) {
+      const errorMessage = err instanceof Error ? err.message : String(err);
+      const isAborted = (err as Error).name === 'AbortError' ||
+                        errorMessage.includes('aborted') ||
+                        errorMessage.includes('No output generated');
+
+      if (isAborted) {
+        // If we have partial content from streaming, save it
+        if (accumulatedText) {
           const partialMessage: ChatMessage = {
             id: generateId(),
             role: 'assistant',
-            content: streamingContent + ' [stopped]',
+            content: accumulatedText + ' [stopped]',
             createdAt: new Date()
           };
           setMessages(prev => [...prev, partialMessage]);

packages/server/package.json

@@ -20,7 +20,7 @@
     "@types/ws": "^8.5.0",
     "typescript": "^5.0.0",
     "vitest": "^2.0.0",
-    "wrangler": "^4.51.0",
+    "wrangler": "^4.59.1",
     "ws": "^8.18.0"
   },
   "dependencies": {