# Open Interpreter — Run Code via Natural Language

> Run code locally through natural language. Supports Python, JavaScript, Shell and more with a ChatGPT-like interface on your own machine.

## Install

Save as a script file and run:

## Quick Use

```bash
pip install open-interpreter
interpreter
```

That's it. You now have a ChatGPT-like interface in your terminal that can execute real code on your machine.

---

## Intro

Open Interpreter lets you control your computer through natural language. Ask it to create files, analyze data, edit images, browse the web, or run any code — it writes and executes Python, JavaScript, or Shell commands locally. Unlike ChatGPT's Code Interpreter, there are no file size limits, no timeout restrictions, and full internet access. 62,000+ GitHub stars, the most popular open-source local code execution agent.

**Works with**: Claude Code, GitHub Copilot, OpenAI Codex, Ollama
---

## What You Can Do

- Create and edit photos, videos, PDFs
- Control a Chrome browser for research
- Plot, clean, and analyze large datasets
- Manage files and system settings
- Run any Python/JS/Shell code through conversation

> ⚠️ Open Interpreter asks for your approval before executing each code block. You control what runs.

## Usage

### Terminal
```bash
interpreter
```

### Python
```python
from interpreter import interpreter

interpreter.chat("Plot AAPL and META's normalized stock prices")
```

### Programmatic (non-interactive)
```python
interpreter.chat("Add subtitles to all videos in /videos.")
interpreter.chat("Make the subtitles bigger?")  # Continues context
```

### Streaming
```python
for chunk in interpreter.chat(message, display=False, stream=True):
    print(chunk)
```

## Configuration

### Change model
```bash
interpreter --model gpt-3.5-turbo
interpreter --model claude-2
```

```python
interpreter.llm.model = "gpt-3.5-turbo"
```

### Use local models
```bash
interpreter --local
# or with custom endpoint
interpreter --api_base "http://localhost:1234/v1" --api_key "fake_key"
```

### Customize system message
```python
interpreter.system_message += "Run shell commands with -y so user confirmation isn't needed."
```

### Context window & tokens
```bash
interpreter --local --max_tokens 1000 --context_window 3000
```

### Interactive commands
- `%verbose true/false` — Toggle debug output
- `%reset` — Clear conversation
- `%undo` — Remove last exchange
- `%tokens [prompt]` — Check token usage and cost
- `%help` — Show help

## FastAPI Server

```python
from fastapi import FastAPI
from fastapi.responses import StreamingResponse
from interpreter import interpreter

app = FastAPI()

@app.get("/chat")
def chat_endpoint(message: str):
    def event_stream():
        for result in interpreter.chat(message, stream=True):
            yield f"data: {result}\
\
"
    return StreamingResponse(event_stream(), media_type="text/event-stream")
```

## Safety

- Always reviews code before execution (unless `auto_run=True`)
- Be cautious with file-modifying commands
- Consider running in sandboxed environments (Codespaces, Replit)
- [Safe mode documentation](https://github.com/OpenInterpreter/open-interpreter/blob/main/docs/SAFE_MODE.md) available

---



### FAQ

**Q: What is Open Interpreter?**
A: Run code locally through natural language. Supports Python, JavaScript, Shell and more with a ChatGPT-like interface on your own machine.

**Q: How do I install Open Interpreter?**
A: Check the Quick Use section above for step-by-step installation instructions. Most assets can be set up in under 2 minutes.

## Source & Thanks

> Created by [Killian Lucas](https://github.com/KillianLucas) / [OpenInterpreter](https://github.com/openinterpreter). Licensed under AGPL-3.0.
> [open-interpreter](https://github.com/openinterpreter/open-interpreter) — ⭐ 62,800+

Thanks to Killian Lucas for creating the leading open-source local code execution interface. Not affiliated with OpenAI.

---
Source: https://tokrepo.com/en/workflows/ad8f9d27-28a8-43b9-bd14-fef811eec1fb
Author: Script Depot