Spaces:
Paused
Paused
| title: Cloudflare OS | |
| emoji: 🖥️ | |
| colorFrom: yellow | |
| colorTo: red | |
| sdk: docker | |
| app_port: 8787 | |
| pinned: false | |
| license: apache-2.0 | |
| short_description: Cloudflare OS (v2) running on workerd inside a Space | |
| # Cloudflare OS on a Docker Space | |
| This Space runs [Cloudflare OS](https://github.com/cloudflare/cloudflare-os), the open-source "AI productivity OS" from Cloudflare, entirely inside a Docker Space. The whole stack (workshop backend, router, gatekeepers) runs locally on `workerd` via `wrangler dev`, with the pre-built frontend served as static assets. | |
| ## Notes | |
| - **Ephemeral storage**: this uses wrangler's local simulation for Durable Objects/KV. All accounts, workspaces, and gadgets are lost whenever the Space restarts. Demo purposes only. | |
| - **Bring your own model**: after signing in, add a model in the model picker. Any OpenAI-chat-completions endpoint works via the "Ollama" provider, e.g. the Hugging Face Inference router: | |
| - Provider: `Ollama` | |
| - Model: `deepseek-ai/DeepSeek-V4-Flash-0731:deepinfra` | |
| - API URL: `https://router.huggingface.co/v1` | |
| - API token: your `hf_...` token | |
| - Pinned to upstream commit `e1ab8fb`. | |