AI workflows: custom endpoint for bring-your-own-key (Azure OpenAI EU, OVHcloud AI Endpoints)
Problem
The AI nodes in workflows accept our own API key, but the endpoint is fixed
per provider. We cannot point a key at a regional deployment, so message,
comment and lead text goes to the provider's default (US) endpoint. For EU
customers this adds a transfer outside the EU that zero content retention
does not cover, because the retention promise applies only to what Zernio
stores.
Request
1. A "Custom (OpenAI-compatible)" provider with three fields: base URL,
API key, model or deployment name.
2. For Azure OpenAI: accept the resource endpoint and deployment name, and
send the key in the api-key header.
3. Show the endpoint host in the workflow node, so we can see where the
text is sent.
Why it matters
One setting covers Azure OpenAI in an EU region, OVHcloud AI Endpoints,
Mistral and self-hosted models. EU customers can then keep AI processing
inside the EU and list a single, known provider in their records of
processing. Zernio is an EU company with zero content retention; this
closes the remaining gap.
Log in to comment and vote
Comments1
Miquel Palet López
Nov 16
This is on our roadmap.