Documentation

ArcFlare is an open model hub plus a tiny CLI. Run any model with one command. Full source and issues live on GitHub.

Install

Requires Node.js 18+. No dependencies.

git clone https://github.com/Hakeperty/ArcFlare-Code.git

cd ArcFlare-Code && npm link

Pulls the model if needed, then runs it. With Ollama installed it uses that backend for real local inference; otherwise it drops into a demo chat.

arcflare run qwen2.5

arcflare run <model> [prompt]Run a model (auto-pulls if needed)

arcflare pull <model>Download a model into the local store

arcflare listList installed models

arcflare search <query>Search the model registry

arcflare show <model>Show model details + config

arcflare edit <model>Edit a model's Modelfile (system, params)

arcflare create <name> --from <base>Make a custom model

arcflare cp <src> <dst>Copy an installed model

arcflare rm <model>Remove an installed model

arcflare serveStart the local HTTP API (default :11435)

arcflare psShow the running server and its models

Create your own variant from a base model and a system prompt:

arcflare create captain --from qwen2.5 --system "You are a sea captain."

# Modelfile
FROM qwen2.5
SYSTEM You are a sea captain.
PARAMETER temperature 0.7

Start a small server for apps and scripts:

arcflare serve

GET /api/tagsList installed models

POST /api/pull{ name } — install a model

POST /api/show{ name } — model details

POST /api/create{ name, from, system } — custom model

POST /api/generate{ model, prompt } — a response

DELETE /api/delete{ name } — remove a model

Looking for license details? See Licenses.