autogen-vertexai-memory 0.1.13

VertexAI Memory integration for Autogen agents

autogen-vertexai-memory

VertexAI Memory integration for Autogen agents. Store and retrieve agent memories using Google Cloud's VertexAI Memory service.

Features

• 🧠 Persistent memory storage in VertexAI
• 🔍 Semantic search for memory retrieval
• 🔄 Automatic chat context updates
• ⚡ Async/await support
• 🎯 User-scoped memory isolation

Installation

pip install autogen-vertexai-memory

Prerequisites

• Google Cloud Project with VertexAI API enabled
• Authentication configured (Application Default Credentials recommended)
• VertexAI Memory resource created

gcloud auth application-default login

Quick Start

from autogen_vertexai_memory import VertexaiMemory, VertexaiMemoryConfig
from autogen_core.memory import MemoryContent, MemoryMimeType

# Configure
config = VertexaiMemoryConfig(
    api_resource_name="projects/xxxxxxxxxx/locations/us-central1/reasoningEngines/xxxxxxxxxxxxxxxx",
    project_id="YOUR_PROJECT_ID",
    location="us-central1",
    user_id="user123"
)

memory = VertexaiMemory(config=config)

# Add memory
await memory.add(
    content=MemoryContent(
        content="User prefers Python",
        mime_type=MemoryMimeType.TEXT
    )
)

# Query with semantic search
results = await memory.query(query="programming language preference")

# Get all memories
all_memories = await memory.query(query="")

API Reference

VertexaiMemoryConfig

VertexaiMemoryConfig(
    api_resource_name: str,  # Full VertexAI memory resource name
    project_id: str,         # GCP project ID
    location: str,           # GCP region (e.g., "us-central1")
    user_id: str            # User identifier for memory isolation
)

VertexaiMemory

VertexaiMemory(
    config: VertexaiMemoryConfig | None = None,
    client: Client | None = None
)

Methods:

• add(content, cancellation_token=None) - Add a memory
• query(query="", cancellation_token=None, **kwargs) - Query memories (empty query returns all)
• update_context(model_context) - Update chat context with memories
• clear() - Delete all memories (irreversible)
• close() - Cleanup resources

Examples

Update Chat Context

from autogen_core.model_context import ChatCompletionContext

context = ChatCompletionContext()
result = await memory.update_context(context)

Custom Client

from vertexai import Client

client = Client(project="my-project", location="us-central1")
memory = VertexaiMemory(config=config, client=client)

Development

git clone https://github.com/thelaycon/autogen-vertexai-memory.git
cd autogen-vertexai-memory
poetry install
poetry run pytest

Contributing

Contributions welcome! Please submit a Pull Request.

License

MIT License - see LICENSE file for details.

Support

• 📫 GitHub Issues
• 💬 GitHub Discussions

---

Made with ❤️ for the Autogen community