# How to Start the AutoHedge Interactive REPL: Command-Line Guide

> Start the AutoHedge interactive REPL easily by running the autohedge command. Analyze portfolio hedging with natural language prompts after installation.

- Repository: [Swarms/AutoHedge](https://github.com/The-Swarm-Corporation/AutoHedge)
- Tags: how-to-guide
- Published: 2026-09-09

---

**Run the `autohedge` command in your terminal after installing the package to launch the interactive REPL, which invokes the `run_repl()` function from [`autohedge/cli.py`](https://github.com/The-Swarm-Corporation/AutoHedge/blob/main/autohedge/cli.py) and accepts natural language task prompts for portfolio hedging analysis.**

The AutoHedge interactive REPL provides a command-line interface for analyzing portfolio allocations and hedging strategies through conversational prompts. According to the source code in The-Swarm-Corporation/AutoHedge, starting the REPL requires invoking the package's command-line entry point, which initializes the `AutoHedge` core class and begins processing user inputs. This guide covers the exact commands, prerequisites, and internal mechanisms required to launch and use the REPL effectively.

## Prerequisites for Starting the AutoHedge REPL

### Install the Package from PyPI

Before launching the REPL, install AutoHedge using pip. The package registers the `autohedge` console script that maps to the CLI module.

```bash
pip install -U autohedge

```

### Configure Environment Variables

The REPL requires API keys for external language model services. While the REPL will start without them—emitting warnings about missing configuration—full functionality requires proper environment setup. Create a `.env` file based on the repository's `.env.example` or export variables directly:

```bash
export OPENAI_API_KEY="your-api-key-here"

```

The [`autohedge/env_loader.py`](https://github.com/The-Swarm-Corporation/AutoHedge/blob/main/autohedge/env_loader.py) module handles loading these variables when the REPL initializes.

## Command-Line Methods to Start the REPL

### Using the `autohedge` Command

The primary method to start the AutoHedge interactive REPL is running the entry-point command installed with the package. This invokes `run_repl()` defined in [`autohedge/cli.py`](https://github.com/The-Swarm-Corporation/AutoHedge/blob/main/autohedge/cli.py) (lines 44-46):

```bash
autohedge

```

### Using Python Module Execution

Alternatively, execute the package as a module using Python's `-m` flag. This approach calls [`autohedge/__main__.py`](https://github.com/The-Swarm-Corporation/AutoHedge/blob/main/autohedge/__main__.py), which forwards execution to the same CLI entry point:

```bash
python -m autohedge

```

Both methods instantiate the `AutoHedge` class from [`autohedge/main.py`](https://github.com/The-Swarm-Corporation/AutoHedge/blob/main/autohedge/main.py) and present the interactive prompt.

## Internal Architecture of the REPL Startup

When you execute the startup command, the following sequence occurs according to the source code:

1. **Argument Parsing**: The CLI module in [`autohedge/cli.py`](https://github.com/The-Swarm-Corporation/AutoHedge/blob/main/autohedge/cli.py) builds an `ArgumentParser` to handle command-line options.
2. **REPL Initialization**: The `run_repl()` function (lines 44-46 in [`autohedge/cli.py`](https://github.com/The-Swarm-Corporation/AutoHedge/blob/main/autohedge/cli.py)) displays the welcome banner and enters the input loop.
3. **Core Class Loading**: User inputs are passed to the `AutoHedge` class defined in [`autohedge/main.py`](https://github.com/The-Swarm-Corporation/AutoHedge/blob/main/autohedge/main.py), specifically its `run` method, which processes the task prompts and returns JSON-formatted recommendations.
4. **Environment Setup**: The [`autohedge/env_loader.py`](https://github.com/The-Swarm-Corporation/AutoHedge/blob/main/autohedge/env_loader.py) module loads API keys and configuration variables before processing begins.

## Using the AutoHedge REPL

### Entering Task Prompts

Once started, the REPL displays a welcome banner and accepts natural language prompts for portfolio analysis. For example:

```

> Analyze NVDA for 50k allocation
[dim]Running...[/]
Result
──────
<JSON-formatted recommendation from AutoHedge>

```

### Exiting the Session

Terminate the REPL session by typing any of the following commands followed by Enter:

- `quit`
- `exit`
- `q`

The input handler in [`autohedge/cli.py`](https://github.com/The-Swarm-Corporation/AutoHedge/blob/main/autohedge/cli.py) checks for these specific strings to break the execution loop and return control to your shell.

## Summary

- **Install AutoHedge** via `pip install -U autohedge` to access the `autohedge` command used to start the interactive REPL.
- **Launch methods** include running `autohedge` directly or `python -m autohedge`, both invoking `run_repl()` in [`autohedge/cli.py`](https://github.com/The-Swarm-Corporation/AutoHedge/blob/main/autohedge/cli.py).
- **Configure environment variables** like `OPENAI_API_KEY` in a `.env` file or shell exports before starting (handled by [`autohedge/env_loader.py`](https://github.com/The-Swarm-Corporation/AutoHedge/blob/main/autohedge/env_loader.py)).
- **Exit the session** by typing `quit`, `exit`, or `q` at the prompt.
- The REPL processes commands through the `AutoHedge` class in [`autohedge/main.py`](https://github.com/The-Swarm-Corporation/AutoHedge/blob/main/autohedge/main.py), returning JSON-formatted hedging recommendations for your queries.

## Frequently Asked Questions

### What is the AutoHedge interactive REPL?

The AutoHedge interactive REPL is a command-line interface provided by The-Swarm-Corporation/AutoHedge that accepts natural language prompts for portfolio analysis and hedging strategies. It runs in a continuous loop defined in [`autohedge/cli.py`](https://github.com/The-Swarm-Corporation/AutoHedge/blob/main/autohedge/cli.py), processing user inputs through the `AutoHedge.run()` method and returning structured JSON recommendations.

### Do I need an OpenAI API key to start the REPL?

No, the REPL will start without the `OPENAI_API_KEY` environment variable set, displaying warnings about missing configuration but remaining functional. Set the key in a `.env` file or export it in your shell to enable full language model functionality, as the [`autohedge/env_loader.py`](https://github.com/The-Swarm-Corporation/AutoHedge/blob/main/autohedge/env_loader.py) module attempts to load these variables during initialization.

### How do I exit the AutoHedge REPL safely?

Type `quit`, `exit`, or `q` at the prompt and press Enter. The `run_repl()` function in [`autohedge/cli.py`](https://github.com/The-Swarm-Corporation/AutoHedge/blob/main/autohedge/cli.py) explicitly checks for these specific input strings to terminate the input loop gracefully and return you to the shell without raising errors.

### Can I run the AutoHedge REPL without installing via pip?

Yes, if you clone the repository, you can run `python -m autohedge` from the project root directory. The [`autohedge/__main__.py`](https://github.com/The-Swarm-Corporation/AutoHedge/blob/main/autohedge/__main__.py) file provides this entry point, which calls the same `cli.main()` function used by the installed `autohedge` command, allowing you to start the REPL directly from source.