superbrain-sdk 0.1.1

Premium Python SDK for Superbrain Distributed Memory Fabric

Superbrain Python SDK

Superbrain is a high-performance, distributed RAM fabric designed specifically for AI agents. It allows agents to share gigabytes of context and data at microsecond speeds using a Zero-Blob Architecture.

Instead of passing massive data blobs over JSON or gRPC, agents pass 36-byte memory pointers, allowing multiple agents to read from the same distributed RAM pool without network overhead.

🚀 Key Features

• Zero-Blob Architecture: Pass data by pointer, not by value.
• Microsecond Latency: Distributed RAM speeds for inter-agent communication.
• Secure Fabric: Enterprise-grade mTLS identity and AES-GCM-256 end-to-end encryption.
• Natively Fast: Powered by a high-performance C-bridge.

🛠️ Installation

pip install superbrain-sdk

[!IMPORTANT] This SDK requires the libsuperbrain shared library. Download it from the GitHub Releases and ensure it's in your LD_LIBRARY_PATH or DYLD_LIBRARY_PATH.

💻 Quick Start (Python)

from superbrain import Client

# 1. Initialize the client
client = Client('localhost:50050')

# 2. Allocate a 10MB memory segment
ptr_id = client.allocate(10 * 1024 * 1024)
print(f"Allocated pointer: {ptr_id}")

# 3. Write data to the fabric
data = b"Hello from Python Agent!"
client.write(ptr_id, 0, data)

# 4. Read data back (or from another agent)
read_data = client.read(ptr_id, 0, len(data))
print(f"Read: {read_data.decode('utf-8')}")

# 5. Free memory
client.free(ptr_id)

🔒 Secure Fabric (E2EE)

encryption_key = b"your-32-byte-aes-encryption-key-"
client = Client('localhost:60050', encryption_key=encryption_key)

# Data is now automatically encrypted before leaving the agent!

📄 Documentation

For full API reference, visit the Comprehensive Consumption Guide.

⚖️ License

MIT License. See LICENSE for more details.