Request access

Documentation

Everything you need to call the gateway.

Quickstart

export KEYRELAY_TOKEN="<your team token>"
curl https://api.keyrelayhq.com/v1/models -H "Authorization: Bearer $KEYRELAY_TOKEN"

Endpoints

PathDescription
POST /v1/chat/completionsChat completions (streaming supported)
POST /v1/embeddingsText embeddings
GET /v1/modelsModels available to your token
POST /api/generateOllama relay (open models)

Authentication

Send your team token as a Bearer token. Tokens are issued from the vault console by workspace admins. Never embed tokens in client-side code.

Rate limits

Default limits are returned in the x-ratelimit-* response headers.