ADK Plugin: Agent

Overview

FlotorchADKAgent is the main entry point for creating ADK-compatible agents that integrate with FloTorch’s Gateway. It fetches agent configurations, builds tools, and provides a proxy that supports config synchronization.

---

Setup

from flotorch.adk.agent import FlotorchADKAgent

API_KEY = "<your_api_key>"
BASE_URL = "https://gateway.flotorch.cloud"
AGENT_NAME = "<your_agent_name>"

---

API

FlotorchADKAgent

FlotorchADKAgent(
  agent_name: str,  # agent name configured in console
  enable_memory: bool = False,
  custom_tools: list = None,
  base_url: str = None,
  api_key: str = None,
)

Creates an agent manager that fetches configuration from FloTorch Gateway and builds an ADK-compatible agent.

get_agent() -> LlmAgent

Returns a FloTorch agent (ADK-compatible) with automatic config synchronization.

---

Features

Config Synchronization

The agent automatically reloads its configuration from FloTorch Gateway based on the syncEnabled and syncInterval settings in the agent config.

Tool Integration

• Memory Tools: Set enable_memory=True to automatically include FloTorch Memory tools (preload/search) so the agent can read/write memories via your configured provider. This uses the FloTorch Memory SDK under the hood; ensure a Memory Provider is configured in your Console and see the plugin docs here: FloTorch ADK: Memory.
• Custom Tools: User-defined tools passed via custom_tools

---

Usage Example

# Create agent manager
agent_manager = FlotorchADKAgent(
    agent_name="customer-support",
    custom_tools=[my_custom_tool]
)

# Get the ADK agent
agent = agent_manager.get_agent()

# Use with ADK Runner
from google.adk.runner import Runner
runner = Runner(agent=agent)

---

Notes

• Agent configuration must exist in FloTorch Console