Dateien und Ordner
Upload documents, images and audio files.
skai connects to the systems you already use: as a knowledge source, as an AI provider, for sign-in, as an output channel and through open interfaces for your own connections.
130+ connections · 6 categories · open to your own systems
No connection found. Many systems can still be connected via MCP or the REST API.
Content from these sources goes into the knowledge base so the AI can answer from it.
Upload documents, images and audio files.
Regularly ingest the content of individual web pages.
Ingest entire websites via their sitemap.
Knowledge from Atlassian Confluence (cloud and on-premises).
Files and folders from OneDrive.
Files and folders from Google Drive.
Documents and lists from SharePoint.
Files from a Nextcloud instance.
Answers straight from a PostgreSQL database.
Answers straight from a MySQL or MariaDB database.
Answers straight from an Oracle database.
Emails from an Outlook mailbox.
Emails from any mailbox via IMAP.
Objects from Microsoft Azure Blob Storage.
Articles from a Zendesk Help Center.
Files from a network file server (SMB/CIFS).
Articles from Matrix42.
Technical terms and abbreviations as knowledge.
Static references to external resources.
Free-text entries as knowledge.
Administrators decide which providers and models are available in the admin area.
GPT models from OpenAI.
Claude models from Anthropic.
Azure OpenAI, hosted in the EU.
Gemini models from Google.
Models via Amazon Bedrock.
Models from Mistral in the EU.
AI models from the German provider IONOS.
Locally hosted models via Ollama.
Sign-in and automatic synchronisation of users and groups.
Sign-in via Microsoft 365 and synchronisation of users and groups from Microsoft Entra.
Sign-in with a Google account.
The channels skai uses to talk to your users or customers.
AI chat on your website.
AI on the phone via your SIP phone system.
Answer enquiries directly via WhatsApp.
Create chats as tickets and hand them to the support team.
Send answers via Teams (in workflows).
Send answers via Telegram (in workflows).
Send answers by email (in workflows).
Open ways to connect your own systems and trigger actions.
Third-party systems call skai through an interface.
Access for external calls, with configurable rights and limits.
Connect external tools and data sources to the AI in a standardised way.
Experts and workflows trigger actions in your systems (webhook / HTTP).
Send emails from experts and workflows.
Run commands on servers (via custom tools).
Start processes via webhook, schedule or event.
Embed the chatbot on any website with a snippet.
Embed predefined chat widgets without your own chatbot.
Automatically sync users and groups from Microsoft Entra.
The Model Context Protocol connects a growing number of known services – the following have an official or a community server.
The MCP landscape evolves quickly; availability and maturity of individual servers vary. Whether a given service is usable depends on the respective MCP server and your configuration.
Additional databases and line-of-business applications can often still be connected, for example via MCP, the REST API or custom tools. Talk to us about your use case.
Discuss a connection