Artificial Intelligence1 min read

Build an AI Agent From Scratch in Python: The Agentic Loop

By · Published by Everything Blog

In short

The agentic harness coordinates tool usage by interpreting model requests and executing standard Python functions on its behalf. Because the underlying Chat Completions API is stateless, the custom wrapper must resend the entire conversation history with every subsequent interaction. This iterative loop allows the model to process ongoing tool results to chain tasks together until it reaches a final answer. Setting up this custom environment requires Python 3.15, the uv package manager, and a funded OpenAI API key.

Key points

  • Standalone language models achieve agency only when an external harness bridges them with…: Standalone language models achieve agency only when an external harness bridges them with actionable tools and user inputs.
  • Chaining multiple tool calls in an iterative loop enables models to resolve complex tasks…: Chaining multiple tool calls in an iterative loop enables models to resolve complex tasks using sequential operational outputs.
  • Basic Python proficiencies, such as manipulating dictionaries and handling paths with pat…: Basic Python proficiencies, such as manipulating dictionaries and handling paths with pathlib, are sufficient to implement this architecture.

Sources checked by our analysis pipeline: 1

You know how LLMs work, and you may already use AI agents to help you write code. However, an LLM isn’t an agent. What sits between them is the agentic loop. In this tutorial, you’ll build an AI agent from scratch in Python by writing the harness that turns a language model into an agent. You’ll do this in about ninety lines of code, with no frameworks.

By the end of this tutorial, you’ll understand that:

- An AI agent is a language model wrapped in a harness, which is a loop that passes information between you, the model, and a set of tools.

- The Chat Completions API doesn’t remember earlier messages between calls, so the harness resends the whole conversation context with every request.

- A tool is any action the harness can perform for the model, such as running a Python function. The model learns about the tool only through its description, which includes a JSON schema for its arguments.

- The model never runs a tool itself. It returns a tool call, and the harness decides…

Artificial Intelligence