Your AI remembers the conversation you're in

Reopen a conversation on a different device, after a reload, or a week later, and your AI picks up where you left off.

Services now show you the actual offer

Opening a service used to show a title, a description and three generic pitch lines … a fraction of what the creator had actually described.

It now shows the following:

• What's included — the full roadmap, step by step, with the number of sessions

• What you'll get — written by the creator, not boilerplate

• Skill tags and category, delivery format and turnaround — you know what to expect

• The price — posted clearly

• The cover art, the creator's own words about the service, and why their clients choose it

You can ask before you buy. There's now an "Ask <creator>'s AI" entry point right on the service page… so you can get a question answered about what the service actually involves before committing to anything.

This applies both in the app and on the public page anyone can open without signing in.

Your AI recalls things faster

When your AI searches its own memory to answer you, the answer surfaces quicker.

Files you upload stay put

Attaching several files at once used to make them disappear. Your AI no longer loses track of them.

You can attach as many files as you like. Chat used to stop at 10 per drop and silently discard the rest.

A file you send now starts a conversation. Sending a file with no message got you silence. Your AI had nothing to reply to, so it said nothing, and you couldn't tell the file had arrived. Now it responds.

"Save to Knowledge Hub?" appears after your AI replies, not before your own message.

Photos work properly

A large photo no longer breaks the chat. Large images are now resized automatically.

iPhone photos work. HEIC files, what your iPhone camera saves by default, is accepted and readable.

Your AI can help you use uare.ai

Ask your AI how something works (how to build your Human Life Model, how to publish a Thought, how to get started) and it now answers from our actual help center instead of guessing.

Deleting things from your AI's memory

If you asked your AI to forget something and it kept coming back, there's a good chance an uploaded file was the reason. Chat can't remove content that lives in a document (the fix is to make the file private or delete it in the Knowledge Hub) but your AI didn't know that, so it deleted everything it could reach, told you it was done, and the document quietly put it back.

Your AI now names the file, tells you if visitors can see it, and won't tell you something is fully removed when it isn’t.

If you were invited to an organization

If you have an invitation to an organization that never worked, it is working now.

Your organization's messages are separate from your personal ones

A conversation belongs to the space you started it in. Talking to your personal AI while you're in an organization is a normal thing to do, and it doesn't move the conversation.

Smaller fixes

A valid PDF is no longer called corrupt. A problem on our side was being reported to you as "this file is corrupted, check it opens on your device."

A file that can't be processed now says so instead of appearing to have uploaded successfully.

Your AI stops answering its own questions. When it showed you a set of choices it was supposed to wait for your click. Sometimes it didn’t, but now it always does.

Your AI doesn't introduce itself to you. If you're its owner, it knows that.