use-freellmapi

Point a project at FreeLLMAPI — a local proxy that aggregates the free tiers of ~30 LLM providers (Google, Groq, Cerebras, NVIDIA, Mistral, OpenRouter, GitHub Models, Cohere, Cloudflare, Z.ai, ModelScope, Ollama, Kilo, OVH, AI Horde, and more) behind a single endpoint — roughly 4B tokens/month across 251 model families / 358 endpoints. A router picks the best available model per request and fails over when one is rate-limited. The payoff: prototype against real models for free, with zero code changes beyond a base URL and a key.

Don't hardcode the roster or the numbers into advice. Providers come and go, the catalog self-updates from a signed feed, and free installs run on a 30-day catalog trail. Read the live /v1/models and the dashboard rather than trusting any list — including this one.

Tiering note: a FreeLLMAPI project is the toolkit's explicit aggregating multi-provider setup. The model & effort tiering policy (model-adaptation, Model & effort tiering) is provider-relative — within a normal build you pick tiers inside ONE provider's ladder and require explicit provider/privacy/budget consent for cross-provider review — and FreeLLMAPI is its explicit carve-out: here the aggregated free tiers ARE the ladder, and the scarce resource is rate/quota headroom, not dollars.

The whole integration, in three facts