Gateway
LiteLLM
Manage model access and virtual keys. API clients use /v1.
Observability
Langfuse
Review traces, token usage, and estimated model costs.
Open usage dashboard →Decisions
Laya
Get advisory tier suggestions or ask general typed questions from an agent.
Open decision API docs →Model playground
Try a Bedrock model
Use a LiteLLM virtual key. This field stays in this browser tab and is never saved here.
Global processing Kimi may process prompts outside Ireland. Use approved data only.
Laya playground
Check a model tier
A suggestion only. Laya runs on this server and does not call Bedrock.
Agents can also call /laya/decide with their own typed questions.