better-mem0-mcp 1.1.0b19__tar.gz → 1.1.0b20__tar.gz

{better_mem0_mcp-1.1.0b19 → better_mem0_mcp-1.1.0b20}/CHANGELOG.md

@@ -1,3 +1,10 @@
+# [1.1.0-beta.20](https://github.com/n24q02m/better-mem0-mcp/compare/v1.1.0-beta.19...v1.1.0-beta.20) (2026-01-05)
+
+
+### Bug Fixes
+
+* correct search response parsing and improve documentation ([d4d4753](https://github.com/n24q02m/better-mem0-mcp/commit/d4d4753a7d0d3c8eb23ef15c051a7bcb1b000310))
+
 # [1.1.0-beta.19](https://github.com/n24q02m/better-mem0-mcp/compare/v1.1.0-beta.18...v1.1.0-beta.19) (2026-01-05)
 
 

{better_mem0_mcp-1.1.0b19 → better_mem0_mcp-1.1.0b20}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: better-mem0-mcp
-Version: 1.1.0b19
+Version: 1.1.0b20
 Summary: Zero-setup MCP Server for AI memory - works with Neon/Supabase
 Project-URL: Homepage, https://github.com/n24q02m/better-mem0-mcp
 Project-URL: Repository, https://github.com/n24q02m/better-mem0-mcp.git

{better_mem0_mcp-1.1.0b19 → better_mem0_mcp-1.1.0b20}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "better-mem0-mcp"
-version = "1.1.0-beta.19"
+version = "1.1.0-beta.20"
 description = "Zero-setup MCP Server for AI memory - works with Neon/Supabase"
 readme = "README.md"
 license = { text = "MIT" }

{better_mem0_mcp-1.1.0b19 → better_mem0_mcp-1.1.0b20}/src/better_mem0_mcp/docs/memory.md

@@ -17,6 +17,8 @@ Save information to long-term memory. Mem0 automatically:
 {"action": "add", "content": "User prefers dark mode and uses FastAPI"}
 ```
 
+**Response:** `Saved: {results: [{id, memory, event}]}`
+
 ### search
 Semantic search across stored memories. Combines vector search with graph context.
 
@@ -24,6 +26,8 @@ Semantic search across stored memories. Combines vector search with graph contex
 {"action": "search", "query": "coding preferences", "limit": 5}
 ```
 
+**Response:** Formatted list of matching memories with optional graph context.
+
 ### list
 Get all stored memories for a user.
 
@@ -31,6 +35,8 @@ Get all stored memories for a user.
 {"action": "list"}
 ```
 
+**Response:** `Memories (N): - [id] memory text`
+
 ### delete
 Remove a memory by ID.
 
@@ -38,10 +44,17 @@ Remove a memory by ID.
 {"action": "delete", "memory_id": "abc12345-..."}
 ```
 
+**Response:** `Deleted: {memory_id}`
+
 ## Parameters
 - `action` - Required: add, search, list, delete
 - `content` - Required for add: information to remember
 - `query` - Required for search: what to search for
 - `memory_id` - Required for delete: ID of memory to remove
 - `limit` - Optional for search: max results (default: 5)
-- `user_id` - Optional: scope memories to a specific user
+- `user_id` - Optional: scope memories to a specific user (default: "default")
+
+## Technical Notes
+- Mem0 API returns `{"results": [...]}` for both `search()` and `get_all()`
+- Graph context is automatically included in search results when available
+- Embeddings use 1536 dimensions for pgvector HNSW compatibility

{better_mem0_mcp-1.1.0b19 → better_mem0_mcp-1.1.0b20}/src/better_mem0_mcp/server.py

@@ -108,10 +108,10 @@ async def memory(
             if not query:
                 return "Error: 'query' required for search action"
 
-            # Vector search - returns {"memories": [...]}
+            # Vector search - returns {"results": [...]} (same as get_all)
             response = _memory.search(query, user_id=user_id, limit=limit)
             memories = (
-                response.get("memories", []) if isinstance(response, dict) else response
+                response.get("results", []) if isinstance(response, dict) else response
             )
 
             # Add graph context