FreeLLM

The simplest Python client for free access to top-tier AI models via public endpoint

freellm is a lightweight, easy-to-use Python package that gives you instant access to powerful models like GPT-4.1 Nano, DeepSeek, Gemini Flash Lite, and Claude 3 Haiku — completely free, no API key, no registration required.

It works by communicating directly with the public web interface, delivering high-quality responses with perfect formatting and minimal setup.

Features

• Zero setup — no accounts, no keys
• Simple .ask("your message") interface
• Four powerful models: gpt (default), deepseek, google, claude
• Optional conversation memory (sends full history when limit is enabled)
• Per-conversation message limit with automatic reset
• Streaming support (token-by-token output)
• Perfect handling of newlines and spacing (no stuck words or visible \n)
• Clean and intuitive CLI
• Minimal dependencies (only requests)

Installation

pip install freellm

Requires Python 3.8+

Quick Start

Programmatic Use

from freellm import FreeLLM

# One-shot query (GPT-4.1 Nano by default)
print(FreeLLM().ask("Tell me a joke"))

# Use DeepSeek
print(FreeLLM(model="deepseek").ask("Explain quantum computing in simple terms"))

# With memory + limit
bot = FreeLLM(model="claude", limit=20)
bot.ask("My name is Alice")
print(bot.ask("What is my name?"))

Interactive Chat (CLI)

freellm                    # GPT-4.1 Nano, no memory
freellm --model deepseek   # Use DeepSeek
freellm --model google     # Gemini 2.0 Flash Lite
freellm --model claude     # Claude 3 Haiku
freellm --limit 15         # Enable memory (up to 15 user messages)
freellm --stream           # Token-by-token streaming
freellm --model deepseek --limit 20 --stream  # All features combined
freellm "Hello, who are you?"  # One-shot message

Usage Examples

# Persistent chat with Claude
bot = FreeLLM(model="claude", limit=10)
bot.ask("Explain how neural networks work")
bot.ask("Now give a real-world analogy")
bot.ask("Make it even simpler for a child")

# Quick stateless queries with different models
questions = ["Capital of Japan?", "Best way to learn Python?", "Write a haiku about rain"]
models = ["gpt", "deepseek", "google"]

for q, m in zip(questions, models):
    print(f"[{m.upper()}]: {FreeLLM(model=m).ask(q)}\n")

CLI Options

freellm --help

usage: freellm [-h] [--model {gpt,deepseek,google,claude}] [--limit LIMIT] [--stream] [message]

FreeLLM - Free access to DeepSeek, Gemini, Claude & GPT 

positional arguments:
  message               Send a single message and exit

options:
  -h, --help            show this help message
  --model {gpt,deepseek,google,claude}
                        Model: gpt (default), deepseek, google, claude
  --limit LIMIT         Enable memory: max user messages before conversation reset
  --stream              Show response token-by-token (streaming)

Important Note on Memory

The underlying service is a free public endpoint and does not officially store conversation state.

When you set --limit or limit=N, FreeLLM sends the full conversation history with every request — this provides the best possible context retention.

Memory works reliably for short-to-medium conversations (up to ~20–30 messages depending on length) and may vary slightly with server load.

Author

IMApurbo
GitHub: @IMApurbo

License

MIT License

---

Enjoy frontier-level AI models for free — no barriers, no costs! 🚀
Made with ❤️ by IMApurbo