GravixLayer Memory provides full API compatibility with Mem0 while adding enhanced features like dynamic configuration and multi-index support.
Default Configuration: Uses baai/bge-large-en-v1.5 embedding model, gravixlayer_memories index, AWS us-east-1. Available models: baai/bge-large-en-v1.5, microsoft/multilingual-e5-large, nomic-ai/nomic-embed-text-v1.5
Basic Mem0 Compatibility
Python SDK
JavaScript SDK
from gravixlayer import GravixLayer
client = GravixLayer()
memory = client.memory
# Add memory (same as Mem0)
result = memory.add("I love pizza", user_id="alice")
print(f"Added: {result['results'][0]['memory']}")
# Search memories (same as Mem0)
results = memory.search("pizza", user_id="alice")
print(f"Found {len(results['results'])} memories")
# Get all memories (same as Mem0)
all_memories = memory.get_all(user_id="alice")
print(f"Total: {len(all_memories['results'])} memories")
# Update memory (same as Mem0)
memory_id = result['results'][0]['id']
memory.update(memory_id, "alice", "I absolutely love pizza")
# Delete memory (same as Mem0)
memory.delete(memory_id, "alice")
Migration from Mem0
Before (Mem0)
After (GravixLayer)
from mem0 import Memory
memory = Memory()
result = memory.add("I love pizza", user_id="alice")
results = memory.search("food", user_id="alice")
Key Features
Core Methods (Same as Mem0)
add() - Add memoriessearch() - Search memoriesget_all() - List all memoriesupdate() - Update memoriesdelete() - Delete memories
Enhanced Methods (New)
list_available_indexes() - List available indexes- Per-operation
index_name parameter for all methods
{
"results": [
{
"id": "memory-id",
"memory": "memory content",
"metadata": {},
"created_at": "timestamp"
}
]
}
Parameters
user_id - User identifier (required)messages - Text or conversation arraymetadata - Additional context (optional)infer - Enable AI inference (optional)index_name - Target specific index (optional, new)
