# Prism

> Prism is a self-hosted adaptive AI gateway. Connect supported AI providers once; Prism gives your applications one API, selects a suitable route, switches around unavailable or exhausted accounts, and keeps cost, access and diagnostics in one place.

- Self-hosted: you run the gateway on your own Linux or macOS machine and control its configuration and operational data.
- One OpenAI-compatible API: /v1/chat/completions, /v1/responses, /v1/embeddings, /v1/rerank, /v1/models.
- Name a model, or leave the choice to a selector: auto, fast, smart, cheap, local.
- Connect provider API keys, and connect subscription accounts through OAuth sign-in.
- Prism Personal is free. Provider usage and infrastructure may still cost money. Prism Enterprise is not available yet.

Every page has a Markdown copy at the same address plus `index.html.md`. The whole site in one file: https://prism.tm/llms-full.txt

## Pages

- [Prism](https://prism.tm/index.html.md): what the gateway does, use cases and editions
- [Blog](https://prism.tm/blog/index.html.md): notes and releases
- [Request access](https://prism.tm/request-access/index.html.md): Prism is not available for download yet; this form is the way to get in touch