Spaces:
Paused
Paused
| title: ASE GLM | |
| emoji: ⬛ | |
| colorFrom: yellow | |
| colorTo: gray | |
| sdk: docker | |
| app_port: 7860 | |
| pinned: false | |
| hf_oauth: true | |
| # `inference-api` is the whole point: it is what lets the Space call the router | |
| # with the signed-in viewer's token, so their account is billed and not ours. | |
| hf_oauth_scopes: | |
| - inference-api | |
| hf_oauth_expiration_minutes: 480 | |
| short_description: A minimal code harness for GLM-5.3, on a character grid. | |
| # ASE · GLM | |
| A minimal Hugging Face code harness. The whole interface is one character | |
| grid rendered to canvas ([gutbrains](https://github.com/Bit-Trading-Company/gutbrains)), | |
| and the backend is FastAPI proxying the Hugging Face inference router. | |
| Models: **GLM-5.3**, GLM-5.3 Flash, DeepSeek V4 Pro, DeepSeek V4 Flash. | |
| ## Try it without an account | |
| Press **see what it does**. It draws charts, a heat map, an animation and a | |
| playable game, entirely in your browser — no model is called and nothing is | |
| spent. It is the same tool the model uses when you ask it for a graph. | |
| ## Billing | |
| Inference is billed to **you**, the viewer — via your Hugging Face sign-in, or | |
| a key you supply. This Space holds no inference token of its own, so a visitor | |
| cannot spend the owner's credits: without a sign-in there is simply nothing to | |
| spend. Signing in grants this Space the `inference-api` scope, and the token is | |
| kept in a signed cookie on the server and used only to call the router on your | |
| behalf. | |
| Requests are rate-limited per viewer to keep one free CPU responsive for | |
| everyone. That limit guards the Space, not a wallet. | |