A GitHub repository that gathers "permanent free" AI APIs—no expiry, no credit card—into a single place, laying out each provider's rate limits and OpenAI-compatible endpoints, is drawing attention. It covers dozens of providers and includes setup details aimed at coding tools and AI agents.
Developer Tools · Open Source Catalog
A Living Directory of LLM APIs You Can Use Free — Permanently
A community-maintained catalog of permanent free-tier LLM APIs — no billing info, no trial-credit clock — listing rate limits, context windows, and OpenAI-compatible endpoints for dozens of providers.
~5,500
GitHub stars accumulated
2
Sections: Provider APIs & Inference providers
100%
OpenAI SDK-compatible endpoints — swap base_url & api_key
Permanent free-tier rate limits, drawn to scale
Requests-per-day (RPD) allowance on the free tier — taller column = more headroom
Cerebras shown as ~1M tokens/day (8K context cap on free); others as requests/day. Units differ per provider — compare within the same metric.
Why developers welcome it
Sharply cuts costs for side & student projects
Accurate, specific rate-limit visibility per provider
Plugs into Cursor, Aider, Claude Code via OpenAI SDK
Auto-generated & updated from a single data source
Caveats to weigh
Regional limits (Gemini may be off in EU/UK/CH)
Prompts may be used to improve models per ToS
Instability when hitting rate limits
"Permanent" depends on the provider — move heavy production loads to paid tiers
Providers in the catalog
Google Gemini · 1M ctx
Groq · LPU
Cerebras · ~2,600 tok/s
Mistral AI · ~1B tok/mo
GitHub Models · 45+
Cloudflare Workers AI · 50+
Hugging Face · 1000s
Cohere
NVIDIA NIM
Z AI (Zhipu)
Kilo Code
LLM7.io
A separate "Trial credits" section organizes time-limited, high-value credits from providers such as Fireworks and Baseten.
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article…