llmgw docs
The manual this gateway's software carries — the same pages
llmgw docs prints and the console's Manual shows.
Using the gateway from code
The official OpenAI and Anthropic SDKs work as they are: only the base URL and the key change. Using llmgw from code shows Python, TypeScript, Go and curl for chat, streaming, tools, embeddings, the Responses API, batches, and what llmgw adds — stored conversations, tags, sessions and memory notes.
API reference: every endpoint, to read and to try
Every document
- The console
- Quick install
- Installing llmgw from nothing
- Adding a node
- Onboarding a user
- Using llmgw from code
- Guardrails
- Knowledge bases
- Choosing what to run
- InfiniBand fabrics and NVLink
- Facility-aware placement: rack power, temperature and cooling
- What llmgw stores, and what leaves the machine
- Configuration reference
- Development: llmgw from source on one machine
- The admin API
- llmgw and Kong
- llmgw and Kubernetes
- Large models: several cards, several nodes
- llmgw
- Changelog
The documents describe llmgw in general, not this installation: they name no address, model or key of it.