Skip to main content
The GravixLayer Memory System lets you store and search information for AI applications. Think of it as a smart database that remembers user preferences and conversations.
Public Preview: Gravix Layer is currently in Public preview. Features are experimental and may have issues or break as ongoing updates to API endpoints and models continue.

​
What is Memory?

Memory stores information about users and conversations:
  • User preferences (“I love pizza”)
  • Conversation history
  • Facts and settings
  • Automatically extract insights from chats

​
AI Processing with infer Parameter

Simple Text: Stored exactly as provided 
memory.add("I love pizza", user_id="alice")  # Stored as: "I love pizza"
Conversations: AI extracts meaningful information from messages when infer=True (default) 
messages = [
    {"role": "user", "content": "I'm looking for a restaurant"},
    {"role": "assistant", "content": "What cuisine do you prefer?"},
    {"role": "user", "content": "I love spicy Thai food"}
]
# AI analyzes the conversation and extracts: "User prefers spicy Thai cuisine"
memory.add(messages, user_id="alice")  # infer=True by default
Raw Storage (infer=False):
  • Stores messages without extracting meaningful information
  • No AI analysis, faster processing
  • Preserves exact conversation content

​
QuickStart

  • Python SDK
  • JavaScript SDK
pip install gravixlayer

​
Core Operations

​
Quick Examples

​
Basic Example

  • Python SDK
  • JavaScript SDK
from gravixlayer import GravixLayer

client = GravixLayer()
memory = client.memory

# Simple text - stored as-is
result = memory.add("I love pizza", user_id="alice")
print(f"Stored: {result['results'][0]['memory']}")

# Conversation - AI extracts insights
conversation = [
    {"role": "user", "content": "I'm looking for a restaurant"},
    {"role": "user", "content": "I love spicy Thai food"}
]
conv_result = memory.add(conversation, user_id="alice")  # infer=True by default
print(f"AI extracted: {conv_result['results'][0]['memory']}")

# Search finds both
results = memory.search("food", user_id="alice")
print(f"Found {len(results['results'])} memories")
Expected Output:
Stored: I love pizza
AI extracted: User prefers spicy Thai cuisine
Found 2 memories

​
Custom Configuration (Optional)

The memory system works great with defaults, but you can customize it for specific needs:

​
Configuration Parameters

The memory system has 4 main settings:
  • embedding_model - Converts text to searchable format
  • inference_model - AI that extracts memories from conversations
  • index_name - Where memories are stored (like folders)
  • cloud_provider - Where data is hosted

​
Configuration Example

  • Python SDK
  • JavaScript SDK
from gravixlayer import GravixLayer

client = GravixLayer()
memory = client.memory

# Configure for multilingual app
memory.switch_configuration(
    embedding_model="microsoft/multilingual-e5-large",
    index_name="user_preferences",
    cloud_provider="GCP",
    region="us-east1" 
)

# Add multilingual memories
memory.add("El usuario prefiere pizza", user_id="alice")
memory.add("L'utilisateur aime le café", user_id="alice")

# Search works across languages
results = memory.search("food preferences", user_id="alice")
print(f"Found {len(results['results'])} memories")
Expected Output:
🔄 Switched embedding model to: microsoft/multilingual-e5-large
🔄 Switched to database: user_preferences
🔄 Switched cloud provider to: GCP
🔄 Switched region to: us-east1
✅ Configuration updated successfully
Found 2 memories

​
Available Embedding Models

ModelLanguages
baai/bge-large-en-v1.5English
microsoft/multilingual-e5-large100+
nomic-ai/nomic-embed-text-v1.5English

​
Available Cloud Providers

ProviderRegionsBest For
AWSus-east-1,
GCPus-east1,
Gravixgl-eu-west1,

​
Setup

Set your API key first:
  • Windows (CMD)
  • Windows (PowerShell)
  • Linux/macOS
set GRAVIXLAYER_API_KEY=your_api_key_here

​
Async Support

GravixLayer supports async/await for Python applications:

​
Basic Async Example

from gravixlayer import AsyncGravixLayer

async def main():
    client = AsyncGravixLayer()
    memory = client.memory
    
    # Add memory
    result = await memory.add("I love pizza", user_id="alice")
    print(f"Added: {result['results'][0]['memory']}")
    
    # Search memories
    results = await memory.search("food", user_id="alice")
    print(f"Found: {results['results'][0]['memory']}")

​
Async with Configuration

from gravixlayer import AsyncGravixLayer

async def async_config_example():
    client = AsyncGravixLayer()
    memory = client.memory
    
    # Switch configuration
    await memory.switch_configuration(
        embedding_model="microsoft/multilingual-e5-large",
        index_name="user_preferences"
    )
    
    # Add multilingual memories
    await memory.add("El usuario prefiere pizza", user_id="alice")
    results = await memory.search("comida", user_id="alice")

​
Async Conversation Processing

from gravixlayer import AsyncGravixLayer

async def process_conversation():
    client = AsyncGravixLayer()
    memory = client.memory
    
    conversation = [
        {"role": "user", "content": "I love sci-fi movies"},
        {"role": "assistant", "content": "Great! I'll remember that."}
    ]
    
    # AI extracts memories from conversation
    result = await memory.add(conversation, user_id="alice", infer=True)
    print(f"Extracted {len(result['results'])} memories")
⌘I