> KudoAI / ai-personas

1,500+ AI personas for LLMs and agents.

It's just a JSON file, so you can use it in any environment.

⚡ Installation

pip install ai-personas

🔌 Usage

import ai_personas

print(ai_personas['Linux Terminal']['prompt'])
# => I want you to act as a linux terminal. I will type commands and you will...

Note: Most type checkers will falsely warn ai_personas is not subscriptable because they are incapable of analyzing runtime behavior (where the module is replaced w/ a dictionary for cleaner, direct access). You can safely suppress such warnings using # type: ignore.

💻 Examples

Find personas by keyword:

def find_personas(keyword):
    return [
        persona for persona, data in ai_personas.items()
            if keyword.lower() in data['prompt'].lower()
    ]

print(find_personas('coach'))
# => ['Interview Preparation Coach', 'Life Coach', ...]

Get prompt for a persona:

def get_prompt(persona):
    return ai_personas[persona]['prompt']

print(get_prompt('Food Critic'))
# => I want you to act as a food critic. I will tell you about a restaurant...

Get random personas:

def random_persona(qty=1):
    import random
    random_personas = random.sample(list(ai_personas), qty)
    return random_personas[0] if qty == 1 else random_personas

print(random_persona())
# => e.g. Reverse Prompt Engineer

print(random_persona(10))
# => e.g. ['Internet Trend & Slang Intelligence', 'Tic-Tac-Toe Game', ...]

Get random prompt:

def random_prompt():
    import random
    return random.choice(list(ai_personas.values()))['prompt']

print(random_prompt())

# e.g. =>
#
# Act as a Node.js Automation Script Developer. You are an expert in creating
# automated scripts using Node.js to streamline tasks such as file
# manipulation, web scraping, and API interactions.
#
# Your task is to:
# - Write efficient Node.js scripts to automate ${taskType}.
# - Ensure the scripts are robust and handle errors gracefully.
# - Use modern JavaScript syntax and best practices.
# ...

Fill variables in template prompts:

def fill_vars_in_template(prompt, vals={}):
    import re
    return re.sub(r'\$\{(.*?)\}', lambda match: vals.get(match.group(1), match.group(0)), prompt)

prompt = personas['Node.js Automation Script Developer']['prompt']
filled_prompt = fill_vars_in_template(prompt, {'taskType': 'web scraping'})

print(filled_prompt)

# =>
# ...
# Your task is to:
# - Write efficient Node.js scripts to automate web scraping.
# ...

Combine prompts:

mega_prompt = f'''
When I start w/ sh: follow prompt A. When I start w/ win: follow prompt B.

Prompt A: {ai_personas['Linux Terminal']['prompt']}

Prompt B: {ai_personas['Windows Terminal']['prompt']}
'''

print(mega_prompt)

# =>
#
# When I start w/ sh: follow prompt A. When I start w/ win: follow prompt B.
#
# Prompt A: I want you to act as a linux terminal...
#
# Prompt B: I want you to act as a Windows Terminal...

Build system prompt:

system_prompt = ai_personas['Study Planner']['prompt']

messages = [
    {'role': 'system', 'content': system_prompt},
    {'role': 'user', 'content': 'Create a weekly study plan for calculus'}
]

Use persona w/ an LLM:

from openai import OpenAI

client = OpenAI()

shell_persona = ai_personas['Linux Terminal']['prompt']
shell_cmd = 'echo "UTC time: $(date -u +%H:%M:%S)"'

response = client.chat.completions.create(
    model='gpt-5.4',
    messages=[
        {'role': 'system', 'content': shell_persona},
        {'role': 'user', 'content': shell_cmd}
    ]
)

print(response.choices[0].message.content)
# e.g. => UTC time: 15:23:42

🏛️ License

🤖 Related

DuckDuckGPT  

Install / Readme

GoogleGPT  

Install / Readme

🤖 ai-personas (Node.js)

Install / Readme / 🔌 API usage

More Python utilities / Report bug / Report vulnerability / Back to top ↑