Loading…
Loading…
Your customers want an assistant inside your product that can actually do things. Building that means solving tenant isolation, per-user scope, credential storage, approval flows and an audit trail before you write a single prompt. That is the part FlozentAi is.
Adding an assistant to a multi-tenant SaaS product means every answer must be scoped to the signed-in user’s tenant, and every action must be attributable. Getting that wrong once is a breach.
One application, with development, staging and production environments. Everything below is scoped to one of them.
Your backend already knows who the user is. Token exchange turns that into an end user inside a host tenant — created automatically on first sign-in.
The provider takes a token callback and a deployment id. Authentication stays on your backend; the component never sees a long-lived secret.
Usually your own API first. The assistant can do exactly what you published and nothing else.
Publish the read actions, run it for a few weeks, and read the traces. Deciding which writes to publish from evidence is both safer and a much easier internal conversation than deciding from a roadmap.
No. Your customers are host tenants inside your workspace and their users are end users. They sign in to your product exactly as they do now.
The React component ships with its own tokens so it does not inherit your page styles by accident, and those tokens are yours to override.
Create a workspace, register an application, and import your first API.