Skip to main content
Version: v1

Backend — get your catalog in

This is Axis 1: the server-to-server feed that pushes your products, categories, and pages into VoiceBot so the assistant can answer about your real store. Everything here speaks one contract — the signed ingest protocol — so the choice below is only about which producer matches your stack. Pick one; it composes with any frontend.

You need a catalog before the widget says anything

The widget answers questions using this data. Until a backend sync runs, there is nothing for it to talk about. Get the catalog in first.

How every producer works​

The shape is the same regardless of language:

  1. Pair once — exchange a one-time code for a per-tenant HMAC shared secret (stored encrypted).
  2. Full snapshot — stream every entity as gzipped NDJSON to establish a baseline.
  3. Deltas — push small signed batches as your data changes, plus a periodic full to reconcile deletes.

The credentials, signing, endpoints, and entity shapes are documented once in the Ingest Protocol Reference. The per-stack pages below are the shortest path to a working sync in your language.

PHP / Laravel​

Available now. A Composer package that maps your Eloquent models to the canonical entity schema and runs the whole lifecycle from artisan. Start at the Laravel Server Sync overview.

Node.js​

Available now. A first-party npm package (@monoverse/voicebot-node) for Express/Nest/Fastify catalogs and headless storefronts. It maps your data to the canonical entity schema with typed builders and runs the whole lifecycle from a VoiceBotClient. Start at the Node Server Sync overview.

Python​

Available from source; PyPI release in progress. A first-party package (monoverse-voicebot) for Django/FastAPI catalogs. Same typed builders and VoiceBotClient lifecycle, Python idioms. Until the PyPI release lands, install it from the repo — the package itself is complete. Start at the Python Server Sync overview.

No-code​

Available now. No server to run a producer on? VoiceBot can ingest your store without you writing any code — a managed crawler reads your public pages, and supported platforms can be pulled read-only. See No-code & read-only alternatives.

Next steps​