Agents

An Agent is a configurable, autonomous AI worker on FlyMy.AI. You define what it does (goal), what tools it can use, and a status. The platform handles execution, scaling, and delivery.

Agent Properties

Property	Required	Description
name	Yes	Human-readable name (string)
goal	Yes	Primary instruction for the agent (string, stored as user_prompt)
tools	Yes	List of integer tool IDs (available_tools)
status	No	One of: draft, initialization_required, active, archived

Memory

Conversation memory is available on the underlying model and does not require explicit configuration.

Creating an Agent

Before creating an agent, you need to add tools from the catalog. Tools are created separately and referenced by their integer IDs.

Step 1: Add Tools

Python

from flymyai import AgentClient

client = AgentClient(api_key="fly-***")

# Browse the tool catalog
catalog = client.tools.available()
for t in catalog:
    print(f"{t.name} ({t.type}): {t.description}")

# Add a tool from the catalog
search_tool = client.tools.create(mcp_tool="tavily")

# Configure the tool if needed
if not search_tool.is_configured:
    search_tool = client.tools.provide_config(
        search_tool.id,
        user_response={"TAVILY_API_KEY": "tvly-***"}
    )

JavaScript

import { AgentClient } from "flymyai-js-client";

const client = new AgentClient({ apiKey: "fly-***" });

// Browse the tool catalog
const catalog = await client.tools.available();
catalog.forEach(t => console.log(`${t.name} (${t.type}): ${t.description}`));

// Add a tool from the catalog
const searchTool = await client.tools.create({ mcpTool: "tavily" });

// Configure the tool if needed
if (!searchTool.isConfigured) {
    await client.tools.provideConfig(searchTool.id, {
        userResponse: { TAVILY_API_KEY: "tvly-***" }
    });
}

Step 2: Create the Agent

Python

agent = client.agents.create(
    name="News Digest",
    goal="Find today's top 5 AI news stories and summarize them.",
    tools=[search_tool.id]
)

print(f"Agent created: {agent.id} (status: {agent.status})")

JavaScript

const agent = await client.agents.create({
    name: "News Digest",
    goal: "Find today's top 5 AI news stories and summarize them.",
    tools: [searchTool.id]
});

console.log(`Agent created: ${agent.id} (status: ${agent.status})`);

Full Example

Python

from flymyai import AgentClient

client = AgentClient(api_key="fly-***")

# Set up tools
tavily = client.tools.create(mcp_tool="tavily")
tavily = client.tools.provide_config(tavily.id, user_response={"TAVILY_API_KEY": "tvly-***"})

browser = client.tools.create(mcp_tool="browser")

# Create the agent with multiple tools
agent = client.agents.create(
    name="Technical SEO Auditor",
    goal="""Perform a technical SEO audit of the given website.
    Check page speed, meta tags, mobile responsiveness, and structured data.
    Produce a scored report with actionable recommendations.""",
    tools=[tavily.id, browser.id],
    status="active"
)

JavaScript

import { AgentClient } from "flymyai-js-client";

const client = new AgentClient({ apiKey: "fly-***" });

// Set up tools
let tavily = await client.tools.create({ mcpTool: "tavily" });
tavily = await client.tools.provideConfig(tavily.id, {
    userResponse: { TAVILY_API_KEY: "tvly-***" }
});

const browser = await client.tools.create({ mcpTool: "browser" });

// Create the agent with multiple tools
const agent = await client.agents.create({
    name: "Technical SEO Auditor",
    goal: `Perform a technical SEO audit of the given website.
    Check page speed, meta tags, mobile responsiveness, and structured data.
    Produce a scored report with actionable recommendations.`,
    tools: [tavily.id, browser.id],
    status: "active"
});

Agent Lifecycle

draft  →  active  →  archived

Status	Description
draft	Being configured. Not yet ready for execution
initialization_required	Tools or configuration need to be finalized
active	Live and ready to accept runs
archived	Deactivated. No new runs can be started

You can transition an agent between statuses by updating it:

# Activate a draft agent
client.agents.update(agent_id=agent.id, status="active")

# Archive an agent you no longer need
client.agents.update(agent_id=agent.id, status="archived")

Management

# List your agents
agents = client.agents.list()

# Update an agent
client.agents.update(
    agent_id=agent.id,
    goal="Updated goal.",
    tools=[tavily.id, browser.id]
)

# Delete an agent
client.agents.delete(agent_id=agent.id)

Next Steps

• Runs -- executing agents and getting results
• Tools -- browse the catalog, configure, and attach tools