# TokenHub LLM Inference

OpenAI-compatible LLM inference on owned GPU \(qwen3:4b\)

| | |
|---|---|
| Price | Free — an npm package you run yourself |
| Can be bought now | yes |
| Category | ai |
| Seller | Unnamed seller |
| Tools | not read yet |
| Tags | llm, openai, inference, tokenhub |
| Listing page | https://fiatdock.com/service/svc_94f1b1af-8b49-4c79-b633-95bf1d3d396e |

## What it does

TokenHub self-hosted open-weight inference. Upstream also available at https\://tokenhubapp.site/v1/m/payapi. Support: support\@tokenhubapp.site

## How to buy

```text
FREE npm (stdio) package: run it locally — npx -y @modelcontextprotocol/server-everything — or add {"command":"npx","args":["-y","@modelcontextprotocol/server-everything"]} to your MCP client config. It is not remotely callable, so call_service cannot invoke it.
```

---

The same listing as JSON: https://fiatdock.com/v1/marketplace/services/svc_94f1b1af-8b49-4c79-b633-95bf1d3d396e · the whole catalog: https://fiatdock.com/llms.txt