Skip to content

What is Flockadoo

Flockadoo is a provider network and orchestration layer for running LLM inference. A central orchestrator (API at core.flockadoo.ai) routes requests to providers that run the Flockadoo Provider Client alongside a local LLM engine (vLLM, vLLM Metal, or llama.cpp). The orchestrator exposes an OpenAI-compatible API, so you can use the same client libraries and prompts you use with OpenAI while your traffic is routed to your own or third-party providers.

  • Orchestrator: Manages consumers, API keys, providers, and model pricing; routes chat/completion requests to healthy providers.
  • Provider client: Lightweight service on each provider machine that reports health, registers models and pricing, and handles inference by forwarding them to the local LLM server.
  • API: OpenAI-compatible endpoints (for example /v1/chat/completions, /v1/completions) so existing tooling and code work with minimal changes.

To call models as a buyer, create a consumer key and send a chat request — see Using the API.

To offer GPU inference, start with Software and hardware requirements, then sign up and create a flock, then install the provider client.

Endpoint-level API docs (parameters, payloads) live in the API reference (OpenAPI 3). These pages are guides.