Independent AI agent library based on the OpenAI SDK

Agentify is a Python library for building AI agents and multi-agent systems. Built on the OpenAI-compatible Chat Completions interface, it supports multiple providers (OpenAI, Azure, DeepSeek, Gemini, Claude) with clear abstractions for memory, tools, and orchestration—no heavy framework lock-in.

• Multi-agent orchestration: Teams, pipelines, hierarchies, and dynamic sub-agent spawning
• Memory service: Pluggable backends (in-memory, SQLite, Redis, Elasticsearch) with policies (TTL, limits, token budgets)
• Tools: @tool decorator for auto-schema generation, or custom tool classes. Built-in file I/O, planning, weather, and more
• MCP Integration: Easy connection to MCP servers via StdIO (local) or SSE/HTTP (remote) to use external tools
• Reasoning models: Configure thinking depth, store chain-of-thought, real-time reasoning logs
• Async & parallel: arun() support with automatic parallel tool and agent execution
• Observability: Callback system for monitoring and debugging
• Advanced capabilities: Dynamic workflows, file/directory operations, complex state management

pip install agentify-core

For optional features:

pip install agentify-core[all]  # Installs all optional dependencies

from agentify import BaseAgent, AgentConfig, MemoryService, MemoryAddress, tool
from agentify.memory.stores import InMemoryStore

# 1. Create a simple tool with @tool decorator
@tool
def get_time() -> dict:
    """Returns the current time."""
    from datetime import datetime
    return {"time": datetime.now().strftime("%H:%M:%S")}

# 2. Create memory service
memory = MemoryService(store=InMemoryStore(), log_enabled=True, max_log_length=100)
addr = MemoryAddress(conversation_id="session_1")

# 3. Create an Agent with the tool
agent = BaseAgent(
    config=AgentConfig(
        name="ReasoningAgent",
        system_prompt="You are a helpful assistant.",
        provider="provider",
        model_name="model",
        reasoning_effort="high",  # optional param:"low", "medium", "high"
        model_kwargs={"max_completion_tokens": 5000} # Pass model-specific params
    ),
    memory=memory,
    memory_address=addr,
    tools=[get_time]  # Add your tools here
)

# 4. Run a conversation
response = agent.run(user_input="What time is it?")

Agentify provides powerful primitives that can be combined to build arbitrarily complex systems:

• BaseAgent: The fundamental unit of work.
• Teams: A group of agents managed by a supervisor.
• Pipelines: A sequence of steps where output passes from one to the next.
• Hierarchies: Tree structures for massive delegation.

Because all flows share the same run() interface, you can build Teams made of Pipelines, Pipelines made of Teams, and deeply nested Hierarchies.

Agentify supports both strict workflows (fixed, pre-defined Pipelines and Hierarchies) and dynamic agentic flows, where a supervisor/router agent decides at runtime which agent, Team or Pipeline to call next.

• Getting Started - Installation and first steps
• Core Concepts - Agents, memory, and tools
• Multi-Agent Systems - Teams, pipelines, and hierarchies
• Advanced Features - Vision, streaming, hooks, and more
• API Reference - Complete API documentation

Check out the examples directory for detailed implementations:

• Single Agent Chatbot
• Multi-Agent Teams
• Sequential Pipelines
• Hierarchical Structures

• Fabian Melchor fabianmp_98@hotmail.com