An AI is only as good as the knowledge it can reach. With Workspace Connect, your snori workspace becomes a reliable memory for ChatGPT, Claude, Perplexity or Gemini: the AI reads your pages and tables and can write results back – always only as far as you allow.
What is the AI connection?
For an AI to talk safely to a data source, it needs a shared standard. That is exactly what the Model Context Protocol (MCP) is – an open, vendor-independent interface through which AI assistants access external systems in a structured way. Instead of laboriously copying content and pasting it into a chat, the AI connects directly and in an orderly fashion with your workspace.
snori provides this connection as Workspace Connect. It works with any AI assistant that supports the open standard – including the common systems like ChatGPT and Claude as well as Perplexity or Gemini. Because the standard is open, you are tied to no vendor and can switch the AI without touching your knowledge base. Workspace Connect is included in the plan and costs nothing extra with snori.
snori and the AI: the two directions
The AI connection works in both directions – snori is the source of knowledge and the target of results.
snori as the source – the AI reads your knowledge
You ask your AI assistant something, and it draws the answer from your pages and smart tables in snori – not from general training knowledge, but from your own, current content. That gives you answers that fit your business, grounded in what really sits in your workspace.
snori as the target – the AI writes results back
The AI can place summaries, drafts or new entries back into snori – as a new page or as a row in a table. This way your knowledge keeps growing instead of trickling away in fleeting chat histories.
How to set up the connection
- In snori, open Workspace Connect and create a connection for this specific workspace.
- You set the rights: may the AI only read, also append, change or delete?
- In your AI assistant you add snori as a connection and enter the access.
- Done: from now on the AI accesses your workspace – within the rules you set.
No new AI model is ever built into snori – you connect the assistant you already use with your knowledge.
Who can access your data
This is where the real value lies. An AI connection in snori can never do more than you allow – the same rights logic as with any access key. You decide per connection: may the AI only read, or also write? See only this one workspace? The AI never chooses its workspace itself – it sees exactly the one you assign it, and no other.
Every access is versioned and can be rolled back. If you revoke the connection, the access is dead at once. Your content stays encrypted throughout in a German data centre certified to ISO 27001. So you give an AI access to your knowledge without giving up control over it.
AI, API or automation – when to use what?
The AI connection is the way when an assistant like ChatGPT or Claude should work directly with your knowledge. If instead a program or script should exchange data, you use the Workspace API. If snori should actively report events outward, webhooks are the right way. All three speak the same governance layer – the rights apply equally everywhere.