OurAI

Latest releases

22 AUG 2026FEATURE

Let OurAI pick the model

Choose Auto and OurAI reads the task in your message and picks the model to match — something fast and cheap for a short message, something stronger for the hard ones. The pill above the answer names the model that replied, and tapping it shows why. Auto is part of Pro and up.

  • Asking for the answer again no longer destroys the old one. You can page between them.
  • The model can now correct and delete what it remembers about you, instead of stacking duplicates.
  • Four new models, and every model is now measured in Danish, so Auto picks one that actually writes it well.
  • Attached documents are read with their tables and equations intact, which means fewer false alarms in Redaction.
22 AUG 2026SECURITY

Conversations and files are encrypted in the database

Messages, what the model remembers about you, and the files you upload are now stored encrypted. Each organisation has its own key, so a copy of the database cannot be read without it.

  • Links in answers are checked against a list of known malicious sites.
09 AUG 2026IMPROVEMENT

Web search reads the pages

Web search no longer fetches headlines alone — it reads the pages and afterwards tells you which ones it read, so you can see what the answer rests on.

  • Five questions about your work pick the models your organisation should use.
  • Suggestions for what to ask when a conversation is empty.
09 AUG 2026IMPROVEMENT

Insights shows more

The overview now shows how much sensitive information is being removed, who in the organisation has got started, and when you will run out of tokens. Topics and keywords are ranked by how much room they actually take up.

28 JUL 2026FEATURE

Redaction catches more, and you have the last word

Organisation names, CVR numbers and clear-text passwords are now removed by default. Dates and ages are made imprecise instead of disappearing, so 28.03.2015 becomes marts 2015 and the reply still makes sense.

  • Keep everything, or keep one type of information, straight from the approval. The button always states what actually gets sent.
  • Error messages now say what happened and what you can do about it.
19 JUL 2026FEATURE

The AI model remembers what you tell it

Tell it something about yourself or your work, and it is remembered and used in later conversations. What is remembered passes through Redaction every time, like everything else you write.

  • Claude Fable 5 and Sonnet 5, GPT-5.6, Grok 4.5 and Kimi K3 have joined.
  • Repeated content is billed at the provider's lower price.
14 JUL 2026FEATURE

Encrypted files, and a model choice that sticks

Documents and images made along the way are now encrypted while they are stored. Switching model mid-conversation sticks for the rest of the conversation.

  • The audit log shows exactly what was sent to the model in each round.
  • Businesses share one pool of tokens that never expires.
04 JUL 2026FEATURE

Documents, images and your own account

You can now have documents written and images generated, shown beside the conversation with versions and download. Individuals have their own account page with subscription, payment, profile and their own rules.

  • Administrators can switch off individual features, including the ones that send data outside Europe, and decide which model new conversations start with.
26 JUN 2026FEATURE

Approval in Redaction

Redaction gained the dialog where you see what was found, correct it, add your own terms and approve, before a single word reaches a model.

  • Web search, and names found by a language model rather than fixed patterns alone.
  • CPR numbers and seven other countries' national IDs are caught.
12 JUN 2026FEATURE

First release

OurAI became available with chat on every major model, Redaction on every message, your own rules enforced in front of every model, and Audit and Insights from day one.