1

pip install open-interpreter

1

interpreter

1
2

export OPENAI_API_KEY=sk-xxx
interpreter

1
2
3

interpreter --local
# 或指定 Ollama 模型
interpreter --model ollama/llama3.3

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22

> Create a Python script that scans for open ports on localhost

I'll create a port scanning script for you.

import socket

def scan_ports(host, start_port, end_port):
    open_ports = []
    for port in range(start_port, end_port + 1):
        sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
        sock.settimeout(1)
        result = sock.connect_ex((host, port))
        if result == 0:
            open_ports.append(port)
        sock.close()
    return open_ports

# Scan common ports
ports = scan_ports('localhost', 1, 1024)
print(f"Open ports: {ports}")

Would you like me to run this code? (y/n)

1
2
3
4
5

> Find all Python files in the current directory that contain "password"

> Create a backup of all .md files in a new folder called "backup"

> Show me the structure of this project

1
2
3
4
5

> Install pandas and matplotlib

> Check disk usage and clean up temp files

> List all running processes using more than 100MB memory

1
2
3
4
5

# 每次執行前確認
interpreter --safe_mode ask

# 自動執行
interpreter --safe_mode off

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11

from interpreter import interpreter

# 基本設定
interpreter.llm.model = "gpt-4o"
interpreter.auto_run = True  # 自動執行程式碼

# 開始對話
interpreter.chat("Create a simple web scraper")

# 或程式化呼叫
result = interpreter.chat("What's 2 + 2?", return_messages=True)

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19

from interpreter import interpreter

# 模型設定
interpreter.llm.model = "gpt-4o"
interpreter.llm.temperature = 0.7
interpreter.llm.api_key = "your-key"

# 執行設定
interpreter.auto_run = False
interpreter.safe_mode = "ask"

# 系統訊息
interpreter.system_message = """
You are a security expert assistant.
Help with security analysis and vulnerability assessment.
"""

# 開始
interpreter.chat()

1
2
3

interpreter.llm.model = "ollama/llama3.3"
interpreter.llm.api_base = "http://localhost:11434"
interpreter.offline = True

1
2

interpreter.llm.model = "claude-3-5-sonnet-20241022"
interpreter.llm.api_key = "your-anthropic-key"

1
2
3
4
5

# 僅允許 Python
interpreter.computer.languages = ["python"]

# 禁用系統命令
interpreter.computer.languages = ["python"]  # 移除 shell

1
2
3
4
5

# 使用 Docker
interpreter --docker

# 或 E2B 沙盒
interpreter --e2b

1

interpreter.auto_run = False  # 每次執行前確認

1
2
3
4
5

> Load the CSV file 'logs.csv' and analyze security events

> Create a chart showing the distribution of event types

> Export the results to an Excel file

1
2
3
4
5

> Check if any ports are exposed on this machine

> Analyze the nginx access logs for suspicious requests

> Generate a security report of the system configuration

1
2
3
4
5

> Every hour, check the website https://example.com and alert if it's down

> Monitor the /var/log/auth.log file for failed login attempts

> Create a cron job to backup the database daily

 1
 2
 3
 4
 5
 6
 7
 8
 9
10

from interpreter import interpreter
from fastapi import FastAPI

app = FastAPI()
interpreter.auto_run = True

@app.post("/chat")
async def chat(message: str):
    response = interpreter.chat(message, return_messages=True)
    return {"response": response[-1]["content"]}

 1
 2
 3
 4
 5
 6
 7
 8
 9
10

tasks = [
    "Check disk space",
    "List recent log errors",
    "Show memory usage"
]

for task in tasks:
    result = interpreter.chat(task, return_messages=True)
    print(f"Task: {task}")
    print(f"Result: {result[-1]['content']}\n")

1
2
3
4
5
6

# 需要額外安裝
pip install open-interpreter[os]

interpreter.computer.mouse.move(100, 200)
interpreter.computer.mouse.click()
interpreter.computer.keyboard.write("Hello")

1

screenshot = interpreter.computer.display.screenshot()

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11

# ~/.config/open-interpreter/config.yaml
llm:
  model: gpt-4o
  temperature: 0.7
  api_key: ${OPENAI_API_KEY}

auto_run: false
safe_mode: ask

system_message: |
  You are a helpful assistant focused on security tasks.  

1

interpreter --config security