Modules
Five modules, each unlocking a coherent slice of the product. Chat is the base and always included; the rest are optional.
At a glance
| Module | Billing | Unlocks |
|---|---|---|
| Chat (base) | Flat | The assistant itself — knowledge base, crawler, widget, custom tools, memory |
| FAQ API | Flat | Mined FAQ clusters and the public FAQ endpoint |
| Product Catalog | Flat | Catalog sync, product search, product cards |
| Escalation | Per seat | Human handoff and the agent Inbox |
| Documentation | Per seat | A public documentation site |
🖼️ [Image] — The Modules grid with Chat "Included" and the other four module cards showing prices and their billing type.
Chat — the base module
Everything the assistant needs to answer a visitor:
- Knowledge base — add, edit, and delete knowledge, plus enrichment (summaries, hypothetical questions, keyword extraction, evaluation queries, and optional contextual retrieval). See Adding knowledge and Enrichment & re-indexing.
- Web crawler — crawl and ingest your site, including headless rendering for React/Vue/Angular pages and optional AI cleanup of crawled content. See Web crawling & sitemaps.
- Widget — embedding on your site, custom branding, and widget authentication. See Embedding the widget and Customizing the widget.
- Custom tools — REST API actions the bot can call mid-conversation. See Custom tools & API actions.
- Cross-session memory — visitor context that survives between sessions. See Cross-session memory.
Chat cannot be removed: without it there's no assistant.
FAQ API
Promptly groups the questions your visitors actually ask into clusters, shows them under Dashboard → FAQ (with a gaps view for questions the bot answered poorly), and exposes the answered ones over an API-key-protected public endpoint so you can render them on your own site.
The clusters are mined from real conversations — this module is not a manual FAQ editor. There's no create/edit/delete list of questions to maintain; you improve the answers by improving the knowledge base.
Product Catalog
Connect a store or product API, map its fields, and let the bot search your products and render them as cards:
- catalog connection, schema detection, and field mapping — see Connecting a catalog and Schema detection & field mapping
- scheduled sync, including incremental delta sync — see Sync & scheduling
- product cards in the widget — see Product cards
- per-tool access rules driven by widget user metadata
Escalation — per seat
Handoff from bot to human: escalation triggers, the agent Inbox, routing and skills, SLA and presence, CSAT, and the bot's ability to offer a handoff on its own. See Escalation settings and Working the Inbox.
Seats are consumed by people who work conversations (the Agent role). Owners and Admins aren't metered.
Documentation — per seat
A public documentation site for your product: versioned pages, knowledge base sync, and search. See Documentation pages.
Seats are consumed by the non-admin users you explicitly grant access to. Owners and Admins aren't metered.