Skip to content
Menu

LLM gateway

Prism

Available

One model name with several vendors behind it.

Read the API reference Contact us

An OpenAI-compatible gateway for large language models. Clients use one model name with several vendors behind it. When a vendor runs out of budget, applies a rate limit or goes down, Prism fails over to the next one mid-request without interrupting the client. The API reference was tested against the live gateway with the client binaries themselves, and it documents anything that does not work.

Who it is for. Teams that call LLMs from their own tools and want one key and one endpoint for every vendor.

Status
Available
Part of
Core platform
Product site
prism.rodmena.co.uk/llms.txt
Tags
OpenAI-compatibleFailoverBearer auth

What it does

  • One model name

    Clients name a model and Prism decides which vendor serves it.

  • Failover mid-request

    If a vendor is out of budget, rate-limited or down, the next takes over without the client noticing.

  • OpenAI-compatible

    Existing clients and SDKs work unchanged apart from the base URL and key.

  • Tested API reference

    Each statement in the reference was tested against the live gateway, and known gaps are documented.

Documentation and access details are on the product site, prism.rodmena.co.uk/llms.txt. For a pilot, an integration or procurement questions, contact us.