Libenter

Conversational LLM prompter

A simple Python CLI tool to send prompts to multiple LLM providers (OpenAI, Anthropic) with language support for English, Hebrew, Arabic, and Farsi.

---

Features

• Supports multiple LLM providers via environment variables.

• Run prompts across all configured LLMs.

• Specify response language using a command-line flag:

  • English (en) – default
  • Hebrew (he)
  • Arabic (ar)
  • Farsi (fa)

• Easy to extend with new LLMs or languages.

---

Installation

1. Clone the repository:

git clone <repo-url>
cd <repo-folder>

2. Install dependencies (if any). Currently, no extra dependencies are required for placeholders. For real API calls, install the provider SDKs:

pip install openai anthropic

---

Environment Variables

Set your API keys and providers:

export LLM_PROVIDERS=openai,anthropic
export OPENAI_API_KEY=your_openai_api_key
export ANTHROPIC_API_KEY=your_anthropic_api_key

LLM_PROVIDERS must be a comma-separated list of supported providers (openai, anthropic).

---

Usage

Run the CLI:

python app.py --lang en

You will be prompted to enter prompts (comma-separated):

Enter prompts (comma-separated): Hello, How are you?

Command-line Options

Flag	Description	Choices	Default
-l, --lang	Response language	en, he, ar, fa	en

---

Example

python app.py --lang he

Input:

Enter prompts (comma-separated): Hello, How are you?

Output:

Prompt: Hello
[OpenAI][he] Response to: ענה בעברית.
Hello
[Anthropic][he] Response to: ענה בעברית.
Hello

Prompt: How are you?
[OpenAI][he] Response to: ענה בעברית.
How are you?
[Anthropic][he] Response to: ענה בעברית.
How are you?

The [OpenAI][he] / [Anthropic][he] indicates the LLM provider and language.

---

Adding More Languages

Add your language instruction to BaseLLM.LANGUAGE_INSTRUCTIONS:

LANGUAGE_INSTRUCTIONS = {
    "en": "Answer in English.",
    "he": "ענה בעברית.",
    "ar": "أجب باللغة العربية.",
    "fa": "به زبان فارسی پاسخ بده.",
    "es": "Responde en español."  # example new language
}

---

Adding More LLM Providers

1. Subclass BaseLLM.
2. Define ENV_VAR for the API key.
3. Implement the generate() method.
4. Add the class to LLM_REGISTRY.

---

License

GNU License