Agent SDK
The agenticstar-platform SDK is an infrastructure SDK for developing custom AI agents in Python that run on the AGENTIC STAR platform. Developers freely design agent logic, while the SDK provides platform infrastructure such as databases, event delivery, memory, storage, RAG, and security.
What is the Agent SDK?
| Item | Value |
|---|---|
| Package | agenticstar-platform |
| Version | v0.5.3 |
| Language | Python 3.12+ |
| Distribution | PyPI |
| Design Philosophy | SDK provides infrastructure only. Agent logic is up to the developer |
Module Structure
The SDK consists of 8 modules. You can selectively install only the modules you need via extras.
| Module | Description | Key Classes |
|---|---|---|
| events | Async event delivery (SSE / Webhook / DB) | EventEmitter, StreamingEvent |
| db | PostgreSQL connection + Azure AD auth | PostgreSQLManager, DataAccess |
| rag | Qdrant vector search + Azure OpenAI Embedding | QdrantManager, EmbeddingGenerator |
| storage | Multi-cloud storage (Azure / S3 / GCS) | AzureBlobStorageClient, S3StorageClient |
| auth | AGENTIC STAR Auth API client | AgenticStarAuthClient |
| memory | Semantic memory (Mem0 + Qdrant) | SemanticMemoryClient |
| security | Content moderation + PII detection | AzureSecurityClient, ContentSafetyValidator |
| common | Common utilities (validation, secret masking) | SecretMasker, validate_identifier |
Quick Start
curl
pip install agenticstar-platform[all]==0.5.3
Python
from agenticstar_platform.events import EventEmitter, EventType
from agenticstar_platform.db import PostgreSQLManager, PostgreSQLConfig, DataAccess
# Database connection
db_manager = PostgreSQLManager(PostgreSQLConfig.from_toml("config.toml"))
da = DataAccess(db_manager)
await da.initialize()
# Event delivery
emitter = EventEmitter(execution_id="exec-001")
await emitter.emit_event(EventType.PHASE_START, message="Starting processing")
Getting Started
| Audience | Steps |
|---|---|
| New development | Build a minimal agent with the Quick Start |
| Migrating existing agents | Understand the module structure with the Architecture Guide, then adopt incrementally |
Guides
- Architecture Guide — SDK module structure and design philosophy
- Deploy Guide — Docker build and platform registration
- Events / Streaming Guide — Implementing SSE event delivery
- Memory Guide — Leveraging semantic memory (Mem0)
- Storage Guide — Multi-cloud object storage
- Security Guide — Content moderation and PII detection
Reference
- SDK API Reference — Complete specifications for all modules, classes, and methods