Is this a library or SDK I install in my project?

No. It's a README documenting free and trial-credit LLM API providers. You read it and use the listed providers' own APIs and SDKs directly.

How current is the model and rate-limit information?

The list is generated by a Python script (`src/pull_available_models.py`) that pulls live data from providers, rather than being maintained by hand, which helps keep it accurate. Rate limits and model availability can still change on the provider's side at any time.

Are all listed providers truly free forever?

No. The repo splits providers into a free-tier section and a trial-credits section; trial credits are one-time or time-limited (e.g., $1 to $30, or expiring after a few months).

Do any providers require verification or data sharing?

Some do. Several providers (e.g., Mistral, NVIDIA NIM, NLP Cloud) require phone number verification, and some free tiers (e.g., Google AI Studio outside certain regions, Mistral's free tier) use your data for training unless you opt out or pay.

What license is this repository under?

The repository's metadata does not specify a license.

Does this list include reverse-engineered or unofficial chatbot APIs?

No. The README explicitly excludes services that aren't legitimate, such as those that reverse-engineer existing chatbots.

Free LLM API Resources: A curated

This repository is a curated list of services that offer free or trial-credit access to LLM inference through an API. It's aimed at developers, hobbyists, and anyone building LLM-powered projects who wants to avoid paying for API usage, at least while prototyping. The list is generated by a Python script (src/pull_available_models.py) rather than edited by hand, so entries stay current with what providers actually expose.

The README is organized into two groups: providers with an ongoing free tier, and providers that give trial credits (usually a small dollar amount or token allotment, often time-limited). For each provider, it lists the available models, rate limits (requests per minute/day, tokens per minute), and any special requirements like phone number verification or opting into data training.

Key features

Free provider list: services like OpenRouter, Google AI Studio, NVIDIA NIM, Mistral, HuggingFace Inference Providers, Groq, Cohere, GitHub Models, and Cloudflare Workers AI, each with a persistent free tier.
Trial credit list: services like Fireworks, Baseten, Nebius, Novita, AI21, Upstage, and others that give one-time or time-boxed credits (often $1 to $30).
Per-model rate limits: request-per-minute, request-per-day, and token-per-minute figures listed for each model where the provider publishes them.
Requirement flags: notes on phone number verification, data-training opt-in, and other conditions attached to free access.
Auto-generated content: the list is produced by a pull script rather than manual edits, reducing the chance of stale or duplicated entries.
Model catalogs per provider: links to each provider's actual model list (e.g., Qwen, Llama, DeepSeek, Gemini, Gemma variants) so you can check what's actually available before signing up.

Ideal use cases

You're prototyping an LLM app or agent and want to test against multiple models without paying, or before committing to a paid provider.
You're comparing open-weight model access (Llama, Qwen, DeepSeek, Gemma) across providers to find the best free rate limits for your use case.
You're a student, hobbyist, or hackathon participant who needs API access without a budget.
You want to build a fallback/routing setup across several free providers to increase effective throughput.

It's not a good fit if you need production-grade reliability or guaranteed uptime: free tiers and trial credits can change or disappear, and rate limits are often too low for real traffic. It's also not useful if you need a specific proprietary model unavailable on any listed free/trial provider, or if you're not willing to deal with the phone verification and data-training conditions some providers impose. This is a reference document, not an SDK or client library, so it won't do any integration work for you.

Installation

There's no software to install to use the list itself: it's a README you read and copy provider links, model names, and API endpoints from directly into your own code.

If you want to run or modify the generator script that produces the list:

git clone https://github.com/cheahjs/free-llm-api-resources.git
cd free-llm-api-resources
pip install -r requirements.txt
python src/pull_available_models.py

This regenerates the README's model tables from each provider's live API. Check the repo's requirements.txt and script for provider-specific API keys or environment variables needed to query each service's model listing endpoint. Since the README is auto-generated, don't edit README.md directly, any changes should go into the generation script.