Integrations

The Integrations catalogue

The Integrations page (sidebar → Integrations) is the marketplace of connectors that feed content into the AI assistant's knowledge base or extend its capabilities.

Layout

Integrations are divided into two sections:

• Installed integrations — connectors you have already configured and connected.
• Available integrations — connectors you can add.

Browsing and filtering

• Search bar — type to filter by integration name or description.
• Tag filters (top-right of the toolbar) — click a tag to show only connectors of that type:
  • 🟠 Knowledge Base — ingest content into the AI's context
  • 🟣 Tool — give the AI new capabilities
  • 🔵 Communication Channel — connect messaging platforms

Connecting a new integration

1. Locate the connector card in the Available integrations grid.
2. Click the card.
3. You are taken to the integration's setup page (only admins can connect integrations).

A maximum of 3 connections per integration type is allowed.

Website crawler integration

The Website integration crawls your website pages and indexes them so the AI can answer questions from the content.

Connecting a website

1. Open Integrations and click the Website card.
2. You are taken to the integration detail page which shows all existing website connections.
3. Click + Neue Verbindung (New Connection) to add a URL.
4. Enter the full URL (e.g. https://example.com) and save.

Monitoring crawl progress

Once connected, the Website detail page (Integrations → Website → [URL]) shows:

Stat	Meaning
Website	The root URL being crawled
Pages	Total pages discovered
Status	Connected, Crawling, or Error

A progress bar shows how many URLs have been processed vs. the total queue. The page updates in real time using live database push notifications.

Starting a sync / re-crawl

Click Synchronisieren (Synchronize) in the top-right corner. The crawler starts from scratch and re-indexes all pages.

Stopping the crawler

While crawling is active the button changes to Stoppen (Stop). Click it to halt the crawler. The system will finish processing any URLs already in the queue before stopping.

Deleting indexed pages

Select rows in the Gecrawlte Seiten (Crawled Pages) table and click the delete action that appears. Deleted pages are removed from the index immediately.

Cloud storage integrations (OneDrive, Google Drive, etc.)

Cloud storage connectors (OneDrive, Google Drive, Dropbox, SharePoint) let you index files and folders directly from your cloud drive.

Stat cards on the connection detail page

Card	Meaning
Files	Number of files discovered, plus how many are indexed
Folders	Number of folders discovered
Indexed	Files whose content has been embedded into the knowledge base

Navigating folders

The file/folder table lets you drill down into sub-folders. The breadcrumb at the top of the page tracks your path. Click any crumb to jump back up.

Syncing

Click Synchronisieren (Synchronize) to trigger a fresh sync of all files in the connected account. The table auto-refreshes during a running sync.

Stopping a sync

If the source is a website/crawler, a Stoppen button replaces the sync button while the crawler is running.

SQL database integration and the Embedding Pipeline

The SQL database integration pulls articles (product data, knowledge records, etc.) from an external SQL database and converts them into searchable embeddings.

Article list

The detail page shows a table of all articles fetched from the connected database. Each row shows the article number, name, and whether its embedding is current.

Stale embeddings warning

If articles have been updated since their last embedding was created, a yellow warning banner appears at the top:

"Es fehlen N Embeddings oder sie sind veraltet."

The Pipeline starten (Start Pipeline) button becomes active once 5 or more articles need updating.

Running the pipeline

1. Click Pipeline starten (N offen).
2. The pipeline status badge changes from idle to running.
3. The table refreshes automatically when the pipeline completes.

Stopping the pipeline

Click Stoppen if you need to halt a running pipeline mid-way.

The pipeline button is disabled if fewer than 5 articles are stale — this threshold prevents unnecessary processing overhead.

Contacting support for integration issues

Each integration detail page has a green Support kontaktieren (Contact Support) button in the left panel.

Steps

1. Go to Integrations and click on the integration that is having problems.
2. On the integration detail page, look at the left panel — you'll see a green info box.
3. Click Support kontaktieren.
4. A support dialog opens where you can describe the issue.

Alternatively, email the ragistry team directly:

• [email protected]
• Or contact the named support people listed on your Profile page.