mfoud444/ollamafreeapi

OllamaFreeAPI – Free Access to Open‑Source LLMs

What it is

• A Python package that wraps a set of publicly hosted endpoints offering a handful of open‑source large language models (LLMs). The library lets you call those models via a simple chat or stream_chat method, without needing to run the models yourself or sign up for a paid service.

Key models

• LLaMA 3.2 (3 b parameters)
• Mistral (latest release)
• DeepSeek‑R1 (built on Qwen)
• Gemma‑based gpt‑oss:20b
• Several smaller or specialized models (e.g., bakllava, smollm2).

How you use it

from ollamafreeapi import OllamaFreeAPI
client = OllamaFreeAPI()
# One‑shot request
resp = client.chat(model="gpt-oss:20b", prompt="Explain neural networks like I'm five")
print(resp)
# Streaming request
for chunk in client.stream_chat('What is quantum computing?', model='llama3.2:3b'):
    print(chunk, end='')

The package handles endpoint discovery, load‑balancing, and optional streaming of tokens.

Features highlighted in the README

• Always free – the service claims no credit‑card or usage limits.
• Model catalog – 16 active models across three families, with 6 server nodes.
• Convenient API – methods to list models, fetch model info, see server locations, and retrieve generic LLM parameters (useful for LangChain integration).
• Documentation – separate markdown files for client reference, examples, model list, and a Termux guide for Android users.

Who might benefit

• Developers or hobbyists who want to experiment with open‑source LLMs quickly, without provisioning hardware.
• Students or educators looking for a zero‑cost way to demo language‑model capabilities.
• Small projects that can tolerate the latency and reliability of community‑run endpoints.

Limitations / considerations

• The service relies on community‑hosted nodes; uptime and performance are not guaranteed.
• No authentication or usage quotas are mentioned, which may affect scalability.
• The README does not detail rate‑limiting, request size limits, or security (e.g., TLS verification).

Installation

pip install ollamafreeapi --upgrade

License – MIT, so you can freely use and modify the client code.

---

All information above is taken directly from the repository’s README; no external assumptions have been added.