T3chatbot.
EN DE
30-Day Free Trial

EXT:AI_CHATBOT · TYPO3 V11–V14 · PHP 7.4–8.4

Turn your TYPO3 Knowledge, Answers, Conversations, Intelligence

Transform your TYPO3 website into an intelligent, 24/7 knowledge hub with multi-provider AI, private Ollama models, and document RAG.

Your content. Your AI. Answers your visitors can actually use.
composer require net2typo/ai-chatbot

Install EXT:ai_chatbot in your TYPO3 project to turn pages and documents into grounded answers.

  • 30-day free trial 
  • Plans from €10/month 
  • Managed Ollama API from €12/month
See pricing
Your knowledge. A conversation away. Captioned product demo · T3chatbot configuration walkthrough

The AI of your choice.
Your knowledge.

Connect your TYPO3 instance to state-of-the-art LLMs. Switch between cloud providers and local Ollama without redesigning the experience.

One knowledge layer. Cloud flexibility or local control.

Flexible embedding providers, including Voyage AI, support retrieval across TYPO3 content and documents.

Two ways to run AI.
One TYPO3 knowledge layer.

CLOUD AI

Launch your assistant and keep your knowledge growing.

Turn pages and documents into grounded answers. Connect your preferred AI provider and give every visitor a clearer path to the information they need.

  • Multi-provider intelligence

    OpenAI, Gemini, Claude, and OpenRouter.

  • Answers grounded in your content

    RAG retrieval across TYPO3 pages and documents.

  • Knowledge that stays up to date

    Auto-indexing, TYPO3 Scheduler, and CLI workflows.

OLLAMA / ON-PREMISE

Your AI.
Your server.
Your data.

Keep inference and embeddings on your own infrastructure. Choose Ollama for a local AI setup managed directly from your TYPO3 backend.

Start with local AI

Five ways your TYPO3 knowledge gets smarter.

Scroll through the capabilities that take content from your CMS to a useful, source-aware conversation.

Capability 01 of 5

Multi-provider AI

Choose the model that fits the moment.

Connect Gemini, OpenAI, Claude, OpenRouter, or Ollama locally or remotely while keeping one TYPO3 knowledge layer.

Capability 02 of 5

Permission-aware document crawler

Every useful source, indexed safely.

Discover pages, RTE links, and FAL attachments, extract PDF, DOCX, DOC, XLSX, CSV, and TXT content, and check frontend user-group restrictions before anything enters the public index.

Capability 03 of 5

Starter plan and up

Built-in input guard

Stop bad input before it costs you.

Every visitor message is checked before it reaches the AI model, blocking prompt injection and spam without spending a single token.

Capability 04 of 5

Duo plan and up

Prompt Studio

Make every answer sound like you.

Create, reorder, and toggle persona, tone, domain-boundary, and output rules from the backend.

Capability 05 of 5

One Click Local Ollama Installation

Your private AI, ready to run.

Install Ollama from your TYPO3 backend, manage models in place, and keep generation and embeddings on your own infrastructure.

Your knowledge.
A conversation away.

Understand every conversation.

See what visitors ask, which sources ground the answer, where knowledge gaps appear, and which conversations are ready for follow-up.

Opt-in, privacy-first analytics: visitors are asked when a chat starts, and only conversations they allow are logged.

Dashboard preview

Starter plan and up

Example data

  • 1,248

    Total queries

  • 87

    Captured leads

  • 342

    Unique sessions

  • 2.85

    Avg. context sources

Chat volume trend Last 14 days

AI topic categories

  • Pricing
  • Technical Support
  • Careers
  • Sales & Demo
  • General Greetings
  • Other

Conversation replays

Session ID Total chats Inspect
session_8f21ac 5 messages Inspect
session_3d9e40 3 messages Inspect
session_c71b02 7 messages Inspect

Website content gaps

  • “Does your core extension support TYPO3 v11?” 34 visitors asked
  • “What is your refund and cancellation policy?” 28 visitors asked
  • “Can I make payments using cryptocurrency?” 19 visitors asked

Safe, private, and fair by design.

Try the widget. Allow or decline analytics, ask a normal question, then try to trick it — and watch the free answers run out.

  1. Ask before tracking

    Visitors choose whether their chat is shared for analytics. The chat works either way.

  2. Check every message first

    The built-in input guard blocks prompt injection and spam before the AI model sees it — no tokens spent.

  3. Upgrade at 100 messages

    After 100 free messages, the site owner is asked to pick a plan.

Free messages left
2
Tokens saved by the guard
0
Analytics
—

Site assistant

example.com · 98 / 100 free messages used

Before we chat

Can this conversation be used for anonymous analytics? It helps the site owner improve answers.

The chatbot works either way.

What just happened

    Events from the demo appear here.

    Built for TYPO3 agencies, enterprise, public sector, and content owners.

    Tailored workflows and architecture whether you manage dozens of client installations or a single mission-critical site.

    Build reusable AI experiences for client sites.

    • Multi-site isolation Site-root tagging keeps every knowledge base separate.
    • Prompt Studio Shape persona, output, and behavior per client.
    • Custom chat UI Make every interaction feel on-brand.
    • Multilingual English and German included.

    Build reusable AI experiences for client sites.

    • Multi-site isolation Site-root tagging keeps every knowledge base separate.
    • Prompt Studio Shape persona, output, and behavior per client.
    • Custom chat UI Make every interaction feel on-brand.
    • Multilingual English and German included.

    Compatible with your TYPO3 stack. Careful with your data.

    Runs on your stack

    Compatibility

    TYPO3 All LTS releases
    1. v11
    2. v12
    3. v13
    4. v14
    PHP 7.4 through 8.4
    Installation
    Composer and non-Composer
    Deployment
    Linux, Apache, and Nginx
    License
    GPL v2+

    Indexes what you publish

    Supported documents

    • PDF
    • DOCX
    • DOC
    • XLSX
    • CSV
    • TXT

    Discovered from

    • TYPO3 public pages
    • FAL attachments
    • Sitemap-discovered content
    • RTE-linked documents

    Privacy & security

    Controls you configure, on infrastructure you choose.

    • Consent first Configurable

      Cloud AI requests can require explicit user consent.

    • Stays on your servers Local

      Local Ollama inference keeps model communication on the customer’s infrastructure.

    • Respects access rules Permission-aware

      Frontend group metadata can protect member-only or restricted files.

    • Isolated per site

      Multi-site embeddings are tagged by site_root ID.

    • Privacy links & disclaimers Configurable

      Configurable privacy links and AI disclaimers are available.

    • GDPR-conscious

      T3chatbot is designed for GDPR-conscious TYPO3 projects.

    These are configurable capabilities and deployment controls — not a substitute for your own legal review.

    From Composer install to your first grounded answer.

    Frequently asked questions

    Answers about providers, documents, permissions, and supported versions.

    Get your TYPO3 AI setup moving.

    Whether you need enterprise onboarding, custom model fine-tuning, or high-volume Ollama hosting.

    Contact Net2typo

    or write to

    • Install

      Composer or non-Composer setup.

    • Configure

      Providers, sites, and permissions.

    • Go local

      Ollama models and private RAG.