Drupal is a registered trademark of Dries Buytaert
drupal 10.6.12 — Update released for Drupal core (10.6.12)! drupal 11.3.13 — Update released for Drupal core (11.3.13)! drupal 10.6.11 — Update released for Drupal core (10.6.11)! drupal 11.3.12 — Update released for Drupal core (11.3.12)! drupal 11.2.14 — Update released for Drupal core (11.2.14)! drupal 10.5.12 — Update released for Drupal core (10.5.12)! cms 2.1.3 — Update released for Drupal core (2.1.3)! drupal 10.5.11 — Update released for Drupal core (10.5.11)! drupal 11.3.11 — Update released for Drupal core (11.3.11)! drupal 11.2.13 — Update released for Drupal core (11.2.13)! drupal 10.6.10 — Update released for Drupal core (10.6.10)! cms 2.1.2 — Update released for Drupal core (2.1.2)! drupal 11.1.10 — Update released for Drupal core (11.1.10)! drupal 10.5.10 — Update released for Drupal core (10.5.10)! drupal 10.4.10 — Update released for Drupal core (10.4.10)! drupal 11.2.12 — Update released for Drupal core (11.2.12)! drupal 11.3.10 — Update released for Drupal core (11.3.10)! drupal 10.6.9 — Update released for Drupal core (10.6.9)! drupal 10.6.8 — Update released for Drupal core (10.6.8)! drupal 11.3.9 — Update released for Drupal core (11.3.9)!
← All modules

multi_role_ai_chatbot

No security coverage
View on drupal.org

Multi-Role AI Chatbot adds a floating, context-aware AI chat assistant to a Drupal site. A single module serves several distinct audiences, each with its own system prompt, welcome message, branding, visibility rules and suggestion buttons.

Audiences

  • Staff console for authenticated team members, with a remote, centrally managed system prompt and access to project and support context.
  • Clients chat for authenticated non-staff users, always scoped to their own data.
  • Anonymous-visitor chat.
  • Unlimited admin-created custom chats (for example "SEO" or "Marketing"), each one configurable independently.

AI or human-staffed chats

Every custom chat can be created as AI-powered or human-staffed. AI chats answer through the configured provider; human chats are attended by real operators from the conversations board and never call the AI. When several chats are visible to a visitor, a chooser lets them pick which one to talk to.

AI providers

  • Talks to Anthropic Claude or Google Gemini.
  • Requests are routed through the shared gateway of the Central Connect AI SDK, which selects the model per tenant and meters usage centrally.
  • Image attachments and lightweight task summarisation are supported on the staff path.

Conversations board and human handover

  • A staff conversations board (inbox) where human-served conversations can be claimed, replied to, released, tagged and archived.
  • Live operator-to-visitor handover: operator replies reach the visitor's floating widget in real time and survive page reloads.
  • @mention handoff: reassign a conversation to another chat by typing @audience — an AI chat co-attends, a human chat takes it over and pulls it into that chat's board.
  • Escalation of technical problems into support task requests when the optional support backend is present.
  • Per-operator read tracking and granular, role-based board permissions (view own, use all tags, assign, purge, learn).

Learning from real conversations

  • A "Learn" workflow distils real conversations into actionable conclusions classified as objections, learnings, gaps or suggestions.
  • Per-chat feedback: each chat keeps its own curated notes and corrections, spliced into that chat's system prompt to improve future answers.
  • Website or product gaps become suggested tasks instead of prompt feedback.
  • Incremental, deduplicated scanning so repeated runs advance through the backlog without re-billing already-seen conversations.

Knowledge and database context

  • Knowledge base: drop text files (TXT, MD, CSV, JSON) per chat; their contents are appended invisibly to the system prompt so the AI can answer from them — visitors never see the raw files.
  • Read-only database context: optionally expose an allow-list of database areas as conversation context, always restricted to what the chat's visibility role may see on the site, and never another user's data.
  • Preset suggestion buttons and per-chat welcome messages.

Branding and appearance

  • Per-audience branding: display name, avatar and brand colour, with inheritance from common defaults.
  • Avatars are Phosphor icons chosen from the shared SDK icon picker; the brand colour is applied across the whole chat interface (header, floating button, tabs, buttons, banners, dropzones).

Budgets, limits and gating

  • Per-chat monthly budget controls over AI usage, with a clear in-chat limit panel and an AI usage meter in the board header.
  • Access gate reflecting the central gateway state (no subscription, credit exhausted, quota exhausted) so paid clients are never blocked by a transient network blip.
  • Entitlement-based gating: the free tier includes one fixed human chat; additional human chats are a paid feature, surfaced with a premium badge and upsell.

Security

  • Hardened, bilingual anti-prompt-injection layer that filters jailbreak and instruction-bypass attempts, plus a response sanitiser that blocks leakage of secrets and internal markers.
  • Default role-based prompting so each audience is answered within its own context, never another user's data.
  • CSRF protection on every state-changing route, rate limiting via the Flood API, and access checks on the board (no IDOR).
  • API keys are read only from settings.php or State, never from exported configuration.

Localisation

English source strings, with a Spanish interface translation shipped in the module and loaded through the bundled .po file.

Standalone and suite integration

The module works standalone. When the optional Aiticketchat module is present, escalated requests are routed into the support task backend and staff sessions can flow from a conversation into a tracked ticket.

Multi-Role AI Chatbot is part of the Central Connect AI suite and uses the shared LLM infrastructure provided by the Central Connect AI SDK.

Activity

Total releases
1
First release
Jun 2026
Latest release
13 hours ago
Release cadence
—
Stability
0% stable

Releases

Version Type Release date
1.0.x-dev Dev Jun 29, 2026