One platform, many client voice agents, no per‑client code

The multi-tenant foundation we start client builds from: onboard a client with configuration, and swap vendors without touching the core.

The problem

Every voice-agent project was starting from zero: sign-in, data, tenancy and the vendor connections. That is where most of the first weeks go.

Each client is a tenant: rows in the database plus validated configuration, with draft and published versions
Row-level security keeps each tenant’s data separate in Postgres
A vendor-free core. Vendors such as Retell, Salesforce, email and telephony are plugs behind clear boundaries, enforced by a lint rule
A single command onboards a new client from a configuration file
Industry playbooks (real estate and dental so far) so a new agent starts from a tested script

Tell us the idea in a few lines. We will say what week one looks like.