Skip to Content

Getting Started

Get up and running with the isA platform in minutes.

Prerequisites

  • Python 3.10+
  • pip or uv package manager
  • (Optional) Docker for local infrastructure

Installation

1. Install the Agent SDK

pip install isa-agent-sdk

Or with uv:

uv add isa-agent-sdk

2. Set Up Authentication

Get your API key from the isA Console and set it. See API Key Authentication for the complete flow:

export ISA_API_KEY="your-api-key"

Or create a .env file:

ISA_API_KEY=your-api-key

3. Verify Installation

from isa_agent_sdk import query async for msg in query("Hello, isA!"): print(msg.content, end="")

Your First Agent

Simple Query

The fastest way to get a response:

from isa_agent_sdk import query async def main(): async for msg in query("What is machine learning?"): print(msg.content, end="") import asyncio asyncio.run(main())

Agent with Tools

Create an agent with access to tools:

from isa_agent_sdk import Agent, ISAAgentOptions agent = Agent( name="research-assistant", options=ISAAgentOptions( allowed_tools=["web_search", "calculator", "code_interpreter"], model="claude-sonnet-4-20250514", system_prompt="You are a helpful research assistant.", ) ) async def main(): result = await agent.run("Find the latest news about AI and summarize it") print(result.text) import asyncio asyncio.run(main())

Or use query() directly with convenience parameters:

from isa_agent_sdk import query async def main(): async for msg in query( "Find the latest news about AI", allowed_tools=["web_search"], ): if msg.is_text: print(msg.content, end="") import asyncio asyncio.run(main())

Agent with Session Memory

Use session_id to maintain context across calls:

from isa_agent_sdk import query # First message async for msg in query("My name is Alice", session_id="conv-001"): print(msg.content, end="") # Later — agent remembers the session async for msg in query("What's my name?", session_id="conv-001"): print(msg.content, end="") # "Your name is Alice"

See Memory & State for persistent backends (Redis, PostgreSQL, Qdrant).

Deployment Options

The steps above use the hosted isA cloud. Prefer to run entirely on your own machine — offline, with no cloud account — or locally with cloud models and team delegation? See Runtime Modes for the local, edge, and cloud options and how to install isA Mate for local execution.

Choose Your Path

I want to build AI agents

  1. Agent SDK Overview - Learn the SDK basics
  2. Tools & Capabilities - Available tools
  3. Memory & State - Persistent memory
  4. Workflows - Multi-step automation

I want to add AI to my app

  1. Model Service - LLM gateway API
  2. MCP Tools - Pre-built integrations
  3. User Services - Auth, payments, storage

I want to deploy to production

  1. Cloud Infrastructure - Kubernetes setup
  2. CI/CD Pipelines - Automated deployment
  3. Monitoring - Observability

Quick Reference

Environment Variables

VariableDescriptionRequired
ISA_API_KEYYour API keyYes
ISA_ENVEnvironment (dev/staging/prod)No
ISA_MODELDefault modelNo
ISA_LOG_LEVELLogging verbosityNo

Available Models

ModelProviderBest For
claude-sonnet-4-20250514AnthropicGeneral use, coding
claude-opus-4-20250514AnthropicComplex reasoning
gpt-4oOpenAIGeneral use
gpt-4o-miniOpenAIFast, affordable
llama-3.1-70bLocalPrivacy, offline

Common Tools

ToolDescription
web_searchSearch the web
calculatorMath operations
code_interpreterRun Python code
file_readRead files
file_writeWrite files
shellExecute commands

Next Steps