Skip to main content

Changelog

Follow new updates and improvements to uare.ai.

November 17th Release - Test Taking Updates, Post to Socials from Anywhere, and more.

A new reader for your Artifacts, with Notes. A new way to take tests, with practice questions. Share to Instagram and LinkedIn from anywhere.

A new reader for Artifacts

  • A cleaner reader. The top bar holds Listen and Aa, where you set the text size, the font, and the page and line spacing. The contents open as a card, and each chapter ends with a card for the next one.

  • Listen starts right away, and picks up where you closed the player. The line being read is highlighted.

  • Notes. Select words to save a passage and add a note to it. Find your passages in the Notes tab, each with its page. They stay in place when the author updates the book.

  • Ask about a passage. Select words to ask the author's AI about them, or to copy them.

  • The author's AI sits beside the book for members, on wider screens.

  • Finished chapters get a tick in the contents.

Courses and tests

  • A new test design. Jump to any question from the grid of numbers, and back again. The magnifier is now a slider that enlarges the whole page. Before you submit, you see which questions are still unanswered or flagged. Results and review are now one page.

  • Practice questions. While reviewing a graded test, ask your teacher's AI for 5, 10 or 20 questions like the one on screen.

  • Better reading passages. Your AI writes English and Reading passages the way the PreACT and the ACT do. Choose which when you ask.

  • For course creators:

    • filter and search a test's questions, or go through them one by one;

    • see how many questions still wait for you;

    • accepting questions is now optional: Publish asks what to do with the rest.

  • Tests imported from a PDF match their answers when the numbering restarts in each section, and when the answer key gives words instead of letters.

  • Lessons that open in order now stay locked in the course until the one before is finished.

  • Teach it Back shows clearly when the microphone is listening.

Share and post

  • Share to your Instagram or LinkedIn from any published post, Artifact, course or service. The share row now shows Instagram and LinkedIn, under the link.

  • The publish card offers every place that can take your post as a link, such as Instagram for a post with no picture.

  • See where each cross-post went, and act on it from there.

  • Your hand edits to an article stay yours. Your AI no longer writes over them.

  • Audiobooks: choose a sound level for each book, none or subtle. Readings use sighs and laughs more sparingly.

Knowledge Hub

  • Refresh is back for your connected social accounts, and brings in only what's new. Disconnect removes the whole connection.

  • If connecting an account fails, its row tells you why.

Chat and your AI

  • Times show in your own time zone, in notifications and everywhere else.

  • Your AI can animate a picture it just made.

  • Pictures you adjust, carousels and character sheets keep their preview in the chat.

  • Course videos over 16 MB upload in pieces, so big videos no longer fail on a slow connection.

  • A page left open during an update reloads itself instead of showing half drawn.

  • A message typed before the chat connects is sent once.

  • Sharing a story or a dream for feedback no longer gets a crisis-helpline reply by mistake.

  • You're told someone chatted with your AI only once they've sent a message.

  • The page of an AI you follow shows Following, and you can unfollow from there.

  • Export my data now includes your knowledge graph, file names and HTML pages, with your messages in order.

  • Search: "Did you mean" appears only for typos or when nothing matches, and searching for AIs matches the topic you typed.

  • An assistant you connect over MCP can use 20 more of your AI's tools. When it wants to do something paid or public, you approve it in that assistant or in the app.

November 13th Release - Desktop - The Biggest Update Yet!

Video and image studios, scheduled tasks, Live Whiteboard lessons, carousel posts, a brandkit and illustrated Artifacts.

Create: Video and Image

Video studio. Turn an idea into a video. Start from a template, such as a course trailer, a product hero, this week's offer or a before and after, or tell your AI what you want. It can star you.

Image studio. Turn an idea into a picture, from templates such as a course cover, a quote card, a product ad, a video thumbnail or a professional headshot.

Carousel posts. Ask your AI for a carousel. It plans 3 to 10 slides in one look and shows one price for the set. You review the slides one at a time and post them as one swipe-able post, or download them as a PDF or as images.

Brand kit. In Settings, then Appearance, pick a typeface, a text colour and an accent colour. The words your AI draws on your pictures and videos use them and stay clear of faces. Organization admins can set one kit for everyone.

More picture shapes. Pictures can be 4:5 or 2:3. Words you want on a picture are drawn exactly, and your AI can add words to a picture or change its shape for free.

Videos ready for each platform. Ask for a Reel and your AI makes the video vertical, at the size platforms expect, with your spoken words as captions.

Voiceover. Add a voiceover to any video, in your AI's own voice, for free, with optional captions.

Redo one shot, or make variants. On a finished video made from a storyboard, you can:

  • Remake a single shot.

  • Try 2 to 10 new openings.

  • Make versions in other languages.

Each one shows its price first.

Remake an ad. Upload an ad and remake it with you, your product, a new place or new words. It keeps the ad's motion and timing.

Other video improvements.

  • Videos of you use the character sheet you approved, as your pictures already do.

  • Videos made of several parts no longer carry the Uare.ai logo.

  • You can make videos in two chats at once.

Courses and Tests

Live Whiteboard, a new course section. Add a video of up to 10 minutes and choose Hand-drawn or 3D simulation. Your AI turns it into a lesson that draws as it plays. Learners can talk to the teacher, who changes the board as it explains, and on a 3D board they can move the model and try things.

Tests

  • A calculator in course tests. It's on by default for new tests, and learners open it from Tools.

  • A quiz retake asks the same question in different words, with the creator's answer choices kept.

  • While you're taking a test, the course's AI stays quiet, and Teach it Back waits without using up a try.

  • Edit and Delete each test from the course builder and from your course page.

  • A test can cover every lesson of a long course.

  • Reading passages in tests. Add a passage by hand or have your AI write one. It shows above its questions.

  • Pictures and tables can go into test questions.

  • A review video for each section of a test, which learners see when they review.

  • Your quiz score updates as each answer is graded.

  • The Level tab explains how points are earned.

Imports

  • Imports keep more questions: maths with absolute-value bars, brackets or root signs, and matching questions.

  • Tests imported from a PDF keep underlines and line numbers.

Course builder

  • Name a lesson section by clicking its label in the builder. Learners see the name as the section's title.

Publishing

General publishing

  • Publishing from a card no longer triggers a paid AI reply.

  • Each hook version of a video keeps its own publish card.

  • A card you haven't acted on survives a reload.

  • You have up to 3 minutes to approve a post your AI prepared.

Instagram

  • Pictures and carousels posted from chat now attach the real pictures.

  • Your AI no longer says a post is live when it was only prepared.

  • When the safety check refuses a post, the publish window says which part to change.

  • Ordinary French, Portuguese, German, Polish and Turkish text is no longer refused by mistake.

Privacy

  • Pictures and videos your AI makes in chat are now private. Only people who can open the chat can see them, until you publish or email them.

Artifacts

Illustrate an Artifact. Your AI agrees a look with you on one sample picture. Then you choose how many pictures, see the price, and each page shows its picture behind the text.

Illustrated chapters

  • Pictures that move with the narration in chapters with both illustrations and an animation.

  • Pages fade into each other over the picture.

  • Watch mode. In illustrated chapters, the picture fills the screen and the spoken line shows as a caption. On a phone, it uses the whole screen.

  • Drag the narration bar to jump anywhere in a chapter.

Opening and sharing

  • A shared Artifact opens on its first picture and says what it holds.

  • An Artifact opens on its page and title right away.

Chat

Sending and delivery

  • Your message appears as soon as you send it. If it can't be delivered, it says so and puts your words and files back in the message box.

  • A message over 64,000 characters is stopped before it's sent, and your text is kept.

  • Typing very fast no longer crashes the chat.

  • Sending the same short reply twice, like "ok", no longer leaves the chat stuck on Stop.

Replies

  • Answers keep coming when the service restarts or your connection drops mid-reply. You no longer see a false "Response interrupted".

  • Leftover tool text, such as facts blocks, no longer shows in replies, including older ones.

  • The floating AI buddy shows search result cards after a search.

Layout

  • New chats and Chat with me open full screen. Open your chat history from the toggle in the chat header.

Paid content

  • Ask an AI about paid content you don't own yet, and it declines politely without giving the content away.

  • Chatting inside an Artifact or course you bought no longer uses up the creator's free messages.

Voice

  • What you say during a brief connection drop is no longer lost.

  • Leaving a voice call's page ends the call and turns off your microphone.

  • If your AI's spoken reply is cut off, it follows with a short question, and you're charged only for the audio you heard.

  • Your AI no longer reads out raw tool tags.

  • Voice note's shortcut: Cmd+Shift+M, or Ctrl+Shift+M on Windows. It works while you're typing.

Your AI's Memory and Files

Memory

  • Answers you give in a Human Learning Model session, typed or spoken, now reliably reach your AI's memory.

  • The Human Life Model is now called the Human Learning Model everywhere. The model picker shows My AI and Pro Mode.

Files

  • Ask your AI about one of your files by its exact name, and it finds that file. If no file has that name, it says so instead of answering from a similar one.

  • A folder's file count includes the files in its subfolders.

  • Deleting a large file removes all of it from what your AI can search.

  • A file you upload right after making a folder goes into that folder. An empty Files panel offers New folder.

  • Links that only lead to an error page are no longer saved to your Knowledge Hub.

Connected Apps

  • Ask your AI to connect an app by its everyday name, for example beehiiv, OneDrive, Teams or Google Maps. You get a one-tap sign-in link for that app.

  • Approving a post once is enough: your AI no longer asks a second time.

  • An assistant you connect to your AI over MCP can now use your AI's own tools on your documents, courses and services. Each use is charged to your credits.

Fixes

  • Renaming a course or editing its description no longer erases it.

  • A course without skills can be published.

  • Reopening a course you're editing no longer makes your AI start changing it.

November 10th Release - Mobile - the Human Learning Model (HLM) and More

New

Use a different account during onboarding. If someone signs in with the wrong email, “Signed in as {email} · Use a different account” now shows on the first onboarding steps. It logs out like Settings does and returns to the welcome screen. Before, the only way out was to finish onboarding as the wrong account.

Skills can be a preference the Skill editor asks “When should this skill be used?”, Always or When relevant. Matches web.

Improvements

Human Life Model is now Human Learning Model everywhere in the app, in all six languages.

Professionals subscription prices must be between $5 and $100 a month.

Skill saves keep the skill file intact, and a skill that failed to load is no longer wiped on save.

Fixes

iOS: reinstalling the app signed you back into the previous account. A fresh install now starts clean, and app updates still keep you logged in.

Sign-up and login keep working once the server stops returning account ids before email verification.

Earlier updates