diff --git a/packages/openai-sdk-python/README.md b/packages/openai-sdk-python/README.md
index 96484503..b0e9c6b2 100644
--- a/packages/openai-sdk-python/README.md
+++ b/packages/openai-sdk-python/README.md
@@ -347,7 +347,7 @@ class SupermemoryOpenAIOptions:
     api_key: Optional[str] = None           # Supermemory API key (or use env var)
     verbose: bool = False                   # Enable detailed logging
     mode: Literal["profile", "query", "full"] = "profile"  # Memory injection mode
-    add_memory: Literal["always", "never"] = "never"       # Auto-save behavior
+    add_memory: Literal["always", "never"] = "always"      # Auto-save behavior
 ```
 
 ### SupermemoryTools
diff --git a/packages/openai-sdk-python/src/supermemory_openai/middleware.py b/packages/openai-sdk-python/src/supermemory_openai/middleware.py
index ea0e1c24..5d748da5 100644
--- a/packages/openai-sdk-python/src/supermemory_openai/middleware.py
+++ b/packages/openai-sdk-python/src/supermemory_openai/middleware.py
@@ -50,8 +50,8 @@ class SupermemoryOpenAIOptions:
     mode: Literal["profile", "query", "full"] = "profile"
     """Memory retrieval mode: 'profile', 'query', or 'full'."""
 
-    add_memory: Literal["always", "never"] = "never"
-    """Memory persistence mode: 'always' or 'never' (default)."""
+    add_memory: Literal["always", "never"] = "always"
+    """Memory persistence mode: 'always' (default) or 'never'."""
 
 
 class SupermemoryProfileSearch:
diff --git a/packages/openai-sdk-python/uv.lock b/packages/openai-sdk-python/uv.lock
index 26688c59..f27d810a 100644
--- a/packages/openai-sdk-python/uv.lock
+++ b/packages/openai-sdk-python/uv.lock
@@ -1,5 +1,5 @@
 version = 1
-revision = 2
+revision = 3
 requires-python = ">=3.8.1"
 resolution-markers = [
     "python_full_version >= '3.10'",
@@ -2427,7 +2427,7 @@ wheels = [
 
 [[package]]
 name = "supermemory-openai-sdk"
-version = "1.0.3"
+version = "2.0.0"
 source = { editable = "." }
 dependencies = [
     { name = "openai" },
diff --git a/packages/tools/README.md b/packages/tools/README.md
index 8c850d6b..5df022e0 100644
--- a/packages/tools/README.md
+++ b/packages/tools/README.md
@@ -298,7 +298,7 @@ interface SupermemoryOpenAIOptions {
   apiKey?: string                // Supermemory API key (or use SUPERMEMORY_API_KEY env var)
   baseUrl?: string               // Custom API endpoint
   mode?: "profile" | "query" | "full"  // Memory search mode (default: "profile")
-  addMemory?: "always" | "never"       // Auto-save conversations (default: "never")
+  addMemory?: "always" | "never"       // Auto-save conversations (default: "always")
   verbose?: boolean              // Enable debug logging (default: false)
 }
 ```
@@ -591,7 +591,7 @@ interface SupermemoryMastraOptions {
   apiKey?: string              // Supermemory API key (or use SUPERMEMORY_API_KEY env var)
   baseUrl?: string             // Custom API endpoint
   mode?: "profile" | "query" | "full"  // Memory search mode (default: "profile")
-  addMemory?: "always" | "never"       // Auto-save conversations (default: "never")
+  addMemory?: "always" | "never"       // Auto-save conversations (default: "always")
   threadId?: string            // Conversation ID for grouping messages
   verbose?: boolean            // Enable debug logging (default: false)
   promptTemplate?: (data: MemoryPromptData) => string  // Custom memory formatting
@@ -665,7 +665,7 @@ interface WithSupermemoryOptions {
 - **conversationId**: Optional conversation ID to group messages into a single document for contextual memory generation
 - **verbose**: Enable detailed logging of memory search and injection process (default: false)
 - **mode**: Memory search mode - "profile" (default), "query", or "full"
-- **addMemory**: Automatic memory storage mode - "always" or "never" (default: "never")
+- **addMemory**: Automatic memory storage mode - "always" (default) or "never"
 
 ## Available Tools
 
diff --git a/packages/tools/src/openai/middleware.ts b/packages/tools/src/openai/middleware.ts
index 45898301..9ceead3c 100644
--- a/packages/tools/src/openai/middleware.ts
+++ b/packages/tools/src/openai/middleware.ts
@@ -34,8 +34,8 @@ export interface SupermemoryOpenAIOptions {
 	mode?: "profile" | "query" | "full"
 	/**
 	 * Memory persistence mode:
-	 * - "always": Automatically save conversations as memories
-	 * - "never": Only retrieve memories, don't store new ones (default)
+	 * - "always": Automatically save conversations as memories (default)
+	 * - "never": Only retrieve memories, don't store new ones
 	 */
 	addMemory?: "always" | "never"
 }
@@ -445,7 +445,7 @@ export function createOpenAIMiddleware(
 		baseUrl,
 		verbose = false,
 		mode = "profile",
-		addMemory = "never",
+		addMemory = "always",
 	} = options
 
 	const logger = createLogger(verbose)
diff --git a/packages/tools/test/openai/unit.test.ts b/packages/tools/test/openai/unit.test.ts
index 18678d01..a8735578 100644
--- a/packages/tools/test/openai/unit.test.ts
+++ b/packages/tools/test/openai/unit.test.ts
@@ -243,6 +243,7 @@ describe("Unit: OpenAI withSupermemory", () => {
 				containerTag: "user-123",
 				customId: "conv-456",
 				mode: "query",
+				addMemory: "never",
 			})
 
 			await wrapped.chat.completions.create({
@@ -435,6 +436,7 @@ describe("Unit: OpenAI withSupermemory", () => {
 			const wrapped = withSupermemory(mockClient, {
 				containerTag: "user-123",
 				customId: "conv-456",
+				addMemory: "never",
 			})
 
 			await wrapped.chat.completions.create({
@@ -448,7 +450,7 @@ describe("Unit: OpenAI withSupermemory", () => {
 			expect(fetchBody.containerTag).toBe("user-123")
 		})
 
-		it("should default to never for addMemory", async () => {
+		it("should default to always for addMemory", async () => {
 			fetchMock.mockResolvedValue({
 				ok: true,
 				json: () => Promise.resolve(createMockProfileResponse()),
@@ -465,8 +467,8 @@ describe("Unit: OpenAI withSupermemory", () => {
 				messages: [{ role: "user", content: "Hello" }],
 			})
 
-			// Only 1 fetch (profile search), no memory add (default is "never")
-			expect(fetchMock).toHaveBeenCalledTimes(1)
+			// 2 fetches: profile search + memory add (default is "always")
+			expect(fetchMock).toHaveBeenCalledTimes(2)
 		})
 	})
 })