Entevra

Search chats, Notifications and project pages

v0.15.452Latest: Search chats, Notifications and project pages

Improved

3 changes
  • Scheduled, project pages and Entevra Workspace line up with the chat’s text column
  • On a phone, Search chats, Notifications, Scheduled and project pages use the message box’s side spacing
  • On a phone, Search chats shares each result’s row between its title and project name, keeping a short project name beside Current

Fixed

3 changes
  • In Notifications, times stay in place when you mark one as read
  • In Notifications, the list scrolls while the heading and actions stay visible
  • Scheduled and loaded project pages no longer repeat Back to chat when the page header already has it

Shared files, downloads and changelog navigation

v0.15.451: Shared files, downloads and changelog navigation

Improved

1 change
  • When you’re signed in, the Entevra ADE download page offers Request access when your account needs an invitation, Check again when no build is available, and Try again when the download check fails

Fixed

6 changes
  • A shared conversation says when Entevra can’t be reached, instead of showing the browser’s network error
  • An image, video or audio file that couldn’t be loaded in a shared conversation offers Try again without reloading the page
  • Try again on a shared file keeps keyboard focus through another failure and moves it to the loaded file only if you haven’t moved to another control
  • If Copy fails for a build’s SHA-256 on the Entevra ADE download page, it explains how to select and copy it yourself
  • In the changelog, the browser’s Back and Forward buttons move keyboard focus to the release they open
  • When Copy fails in Subscribe to the changelog, the card selects the full address and explains how to copy it yourself

Docs search, drop-down lists and image focus

v0.15.450: Docs search, drop-down lists and image focus

Fixed

7 changes
  • Tab and Shift+Tab close a drop-down list and continue from the control that opened it
  • Drop-down lists use their field’s label for screen readers
  • Stop comparing in Voice design returns keyboard focus to Compare on the result
  • The sign-in dialog’s email field keeps its edge when focused, so you can see where you’re typing
  • In the chat and shared conversations, a generated image shows when it has keyboard focus
  • The Docs search button has a name for screen readers on a phone
  • In Docs search, keys used to compose text with an input method no longer move through results or open a guide

Keyboard controls, Reply details and screen readers

v0.15.449: Keyboard controls, Reply details and screen readers

Improved

3 changes
  • A chosen control in Settings, the sidebar or the Files toolbar takes a slightly stronger shade when it has keyboard focus
  • Community’s In reply to link has a larger click target
  • The status page’s footer links have larger click targets on a computer

Fixed

8 changes
  • Tab and Shift+Tab close the account menu and continue from its button
  • The arrow keys in Reasoning effort now reach Less effort and More effort as well as the levels
  • Translation language in Writing tools can be chosen with the keyboard
  • Screen readers announce when Search your skills finds no matches
  • On a computer, Collapse sidebar and Expand sidebar keep keyboard focus on the button
  • On a phone, a project’s expand and collapse control points screen readers to its own list
  • Tab in the Reply details dialog skips controls inside folded sections
  • Upload photo in Settings now shows when it has keyboard focus

The phone sidebar, Skills and Community

v0.15.448: The phone sidebar, Skills and Community

Improved

7 changes
  • On a phone, the sidebar slides in and out where the browser supports it, and opens and closes without motion when reduced motion is on
  • On a phone, the account menu fits within the sidebar’s width, above your account row, with one even shade over the page
  • The Files pane keeps the width you chose when you switch chats, until you reload the page
  • On a phone, the Skills library’s search takes its own row, with sorting and Publisher side by side
  • Community’s discussion list keeps the title and byline on separate lines with less space around each row
  • On a phone or touch screen, the media studio’s examples take more compact rows
  • On a phone, the missing-page buttons fit their labels instead of stretching across the page

Fixed

7 changes
  • On a phone, folders that were already open no longer expand again as the sidebar opens
  • On phones where Back dismisses dialogs, it closes the account menu without closing the sidebar
  • Pressing Enter in Search your skills no longer sends, queues or steers your draft
  • In Skills, a licence stays together where it fits, and the separators no longer begin a new line
  • Search and other fields in Skills keep a clearer edge when focused
  • On a phone, Integrations shows Back to chat once above the Skills library
  • A sign-in method’s badge stays beside its description instead of overlapping it

Room for the model’s name on a phone, and finer replies

v0.15.447: Room for the model’s name on a phone, and finer replies

Improved

4 changes
  • On a phone, the context circle in the message box shows its ring alone, so the model’s name has room. Its reading stays in its label and in Usage in this chat, and the ring keeps any warning colour
  • A reply’s steps are joined by one line from icon to icon
  • A one-line command in a code block takes less height
  • Share uses the same small type as the reply’s other actions

Fixed

1 change
  • In the dark themes, each site’s logo in web search results sits on a light plate, so dark logos no longer disappear

A tighter chat page on a phone, and clearer Settings

v0.15.446: A tighter chat page on a phone, and clearer Settings

Improved

8 changes
  • On a phone, a reply’s actions stay on one row, even beside a version switch, and Share sits closer under the last reply
  • On a phone, replies keep the same edges in short and long chats, and the home’s message box is as wide as a chat’s
  • A file a reply creates has a more compact header, and on a phone a diagram’s header stays on one row
  • A wide equation fades at its edge before you scroll it, where your browser supports it
  • Inline code and source chips stay on a reply’s reading line, and a comma after short inline math no longer starts a new line
  • Words fade out above the button that jumps to the latest reply instead of running under it
  • Settings, in both styles, uses the site’s type sizes, one highlight for the current and chosen items, and tighter rows and gaps
  • On a phone, Settings is centred, and the Queue and Steer choice sits under its explanation

Fixed

4 changes
  • A suggested next message starts at the caret and wraps as the message box does
  • The starter prompt categories keep each dot beside its label on a phone
  • The chat’s chapter rail has a wider click target, without changing how it looks
  • Settings’ content stays aligned when you switch between short and long sections

New menus, the effort card and replies that read your files

v0.15.445: New menus, the effort card and replies that read your files

New

4 changes
  • Replies in a project’s chat can search and read the text files in its Files and folders when your message needs them, and what they read counts toward that reply’s cost. Files you mark Never send to models carry a Never sent badge and are never read. Turn it off with Let replies read your project files in Settings → Personalization
  • Drag the inner edge of the sidebar or the Files pane to resize it, or past its smallest width to close it. Both open at their usual width each time you load the page
  • The home page opens with a short animation: stars write the headline and form the Entevra mark and three tools. It plays once, selecting the mark plays it again, and it stays still with reduced motion or Save-Data
  • RSS feed, on the changelog, opens Subscribe to the changelog: copy the feed’s address or open it in Inoreader

Improved

8 changes
  • The + menu is one short list: Upload or add files (⌘U, or Ctrl+U on Windows), then Recent files, Media, Skills, Connectors and Writing tools as submenus, Starter prompts and Global Radio, then the tools you can use as rows with a check while they’re on
  • The account menu has your email at its head, then Settings (⌘,), Usage, Trash and Community, Read docs and Changelog, and Privacy Policy and Terms of Service, which open in a new tab
  • On a phone, the message box keeps (+), the model and its effort, voice and Send on one row, and the menus, Reply details and Usage in this chat open as a sheet from the bottom of the screen
  • The Reasoning effort card shows the model’s levels on one track: point at a level to read what it does, or step with the arrows. Unless you’ve saved a choice, chats on models with effort levels start at Medium, or the nearest level the model has. A level you pick, or Model default, stays
  • Reply details and Usage in this chat are more compact, and a chat with one reply shows its tokens once
  • A chosen or keyboard-focused item, in the chat’s menus and on the site, now looks like the pointer’s highlight instead of a dark fill
  • The Files and preview pane is more compact, and its Filter files field matches the search field in Skills
  • In the sidebar, a row takes one highlight for hover, selection and focus, and long chat titles fade out at their end where the browser supports it; other browsers keep “…”

Fixed

2 changes
  • Faint words are easier to read: the message box’s placeholder, the numbers and dates on source cards, citation numbers, and Warm’s muted text
  • On a short, wide screen, such as a phone turned sideways, the cookie card fits on one line under the header and no longer covers the home page’s buttons

Code and connectors in long replies, and a faster start

v0.15.444: Code and connectors in long replies, and a faster start

New

2 changes
  • Long replies can now run code once and use your connected services at every step. When the reply continues, its code never runs again
  • Settings sections, Integrations, Media and the chat’s other pages each have their own address, so a link or a reload opens the same place

Improved

8 changes
  • The chat opens faster: the model list is smaller and stays cached between visits, and parts of the page load only when a reply needs them
  • Every page, from the home page to the chat and Settings, now uses one set of type sizes, weights and spacing: replies read at 16px on a 28px line, and titles and short lines sit close to what they introduce
  • The home page has less space between its sections, Pricing less above the footer, and the shared footer tighter spacing above, below and between its links
  • Entevra Sans, our typeface, has round quotes and apostrophes, and figures redrawn to new widths
  • A reply’s steps and the web search pill sit in compact rows, so the answer starts closer to your message
  • The message box, and Community’s post and comment fields, stand out from the page in every theme
  • The web search card lines its sources up along one rail, and opening or closing a source keeps your place
  • On a touch screen, the chat’s small controls and the buttons in Settings, Integrations, the Library, Community and Projects have larger tap targets, without changing how they look

Fixed

4 changes
  • On Android, the message box stays above the keyboard
  • In Safari, double-clicking a figure in a reply no longer greys out the page below it
  • Reloading Community’s Messages, or a conversation in it, opens it again instead of a missing page, and choosing a category leaves Messages
  • A long reply retries a step, or a search inside it, that a busy database delayed. If the wait runs too long, it pauses and keeps the text already written

A new look for the chat, and a warm light theme

v0.15.443: A new look for the chat, and a warm light theme

New

8 changes
  • The chat has a new look: controls without borders, with a soft highlight for hover, keyboard focus and selection, in the chat, the sidebar and the Files pane, and the model with its effort on the right of the message box
  • A circle in the message box shows how much of the context window the chat is using. Select it to open Usage in this chat
  • The context window card shows what fills the latest reply’s context, the free space left and how it is growing. It folds away, like the per-reply card
  • Warm, a new light theme in warm ivory. The themes are now Light, Warm, Dark and Black, plus System, and Light is the default
  • Memory notes are grouped by category in Settings → Memory and on a project’s page, and you can set a note’s category
  • Export notes saves all your memory notes as one Markdown file, and the export of every chat in Settings → Account now includes them
  • Type /plan and a task to get a plan without any changes made. Reply “Go ahead with this plan.” to start it
  • Your own models can name a fallback to answer when they can’t, and Reply details list each model a reply tried and why it didn’t answer

Improved

6 changes
  • The web search, context and price cards fit better on a phone and on a wide screen
  • A scheduled task whose run would cost more than your per-reply cap pauses and tells you why
  • Your own models show their family’s or provider’s mark, or one you choose
  • Web search results show each site’s icon, and numbered sources in a reply become chips that open the page they cite
  • Text to speech reads long texts: they are split into parts and played as one reply with a single download. After a reload, Finish the missing parts continues in the same voice, in the browser that started it, which keeps its settings for up to seven days
  • Compare shows public benchmark scores beside the models, with their source credited

Fixed

1 change
  • A video is charged only once it’s saved. One whose file couldn’t be fetched costs you nothing

Connectors, temporary chats and long replies

v0.15.442: Connectors, temporary chats and long replies

New

6 changes
  • Connectors: connect services you use in Integrations → Connectors, and a reply can read from them when your message is about them. Services that aren’t ready yet offer Notify me, and the bell tells you when one is
  • Temporary chat, in the + menu, stays out of your history and uses no memory. Leaving it deletes it; if that can’t finish right away, it’s cleaned up after 24 hours without a message, and on a suspended account it’s kept until the suspension ends
  • Let Entevra look up your past chats, in Settings → AI behavior: in a chat outside a project, the model can search your other chats outside projects and read parts of them when your question needs it
  • Messages, in Community: a private conversation with Entevra’s administrators about your account, access, billing or anything else. The bell tells you when they reply
  • Files and folders in projects: add and edit text files on a project’s page, with each file’s history and a Trash. Import a folder, a zip or a public GitHub repository, and download everything with Export as zip
  • On supported models, long replies continue on our servers across steps, even after a reload or a closed tab, except in temporary chats. When a reply reaches a limit that can be raised, Keep going shows the price of continuing. The bell tells you when it ends

Improved

4 changes
  • Code execution is open to every member, and files a code run makes appear as cards under the reply
  • Search the web automatically is now on by default and saved to your account. A question about your own items in a connected service is answered from that service, not the web
  • Steer adds your message to a long reply too, at its next step
  • Long replies now cover chats with attached files, images on supported models, Web search, skills, memory and compacted chats

Your own keys, a per-reply cap and Compare

v0.15.441: Your own keys, a per-reply cap and Compare

New

5 changes
  • Per-reply cap in Settings → AI behavior limits what a single reply may cost, including its searches and code runs. It’s $5 by default, within your account’s limits, or a preset or exact amount you choose. A reply that wouldn’t fit isn’t started
  • Use your own provider keys and models: add them in Settings → AI behavior → Your keys, and they appear in the model picker under Custom, marked Yours. Replies on them are billed by your provider, not Entevra
  • Compare in the model picker sets 2 to 4 models side by side, with Copy as table
  • Search chats (⌘K, or Ctrl+K on Windows) adds Find in this chat, the effort level, Export this chat, Memory notes, Usage, the theme and Keyboard shortcuts
  • 168 illustrated avatars to choose from, up from 48

Improved

8 changes
  • Type / in the message box to find commands, tools, media and skills in groups, even from a few letters, on a computer and on a phone
  • Token counts read the same way everywhere, such as 12.4K and 1.45M, and Settings → Usage history shows only your own usage, in a compact card
  • Long chats stay smooth: only the part on screen is drawn, and a streaming reply redraws only itself
  • Automatic chat titles name the topic, and are checked for numbers and dates your message didn’t have
  • Send now sends a queued message with the model, effort and tools it was queued with
  • The goal view marks the step it’s working on, with Working on this now, or Next while the goal waits
  • Your messages show the day as well as the time when they weren’t sent today, as replies do
  • The project’s name in “New chat in …” and before a project chat’s title opens the project, and every full page you reach from the sidebar has Back to chat

Fixed

2 changes
  • Search the web automatically works again for questions in Arabic
  • Keyboard focus no longer jumps to the top of the page after a menu closes, and screen readers hear confirmations, the bell’s unread count and what the model search finds

Also

2 changes
  • The desktop app is now called Entevra ADE
  • The Pricing page: Entevra is free during the public Beta, and credit plans priced from each model’s rates are coming soon

Media Library, picture editing and video from an image

v0.15.440: Media Library, picture editing and video from an image

New

8 changes
  • Media → Library gathers every image, video and voice clip you’ve made across your chats. Open one to view or play it, Download it, or Go to chat
  • Edit a picture and see the original beside the edit. On a picture you generated, Vary makes another version, Reuse prompt fills the message box again, and Keep marks it for the Library’s Kept filter
  • A video can start from an image you upload or one made in the same chat
  • Automatic, in the Video tab, picks a model for your settings and says what it picked and why
  • Voice design makes three voices from one description, and A/B listen switches between two voices at the same point in the clip
  • Each media tab offers up to 12 examples that fill in the model, settings and prompt. Nothing is sent until you choose Generate
  • Export video records an animated SVG preview as a video, in your browser
  • The sidebar has Media and Integrations groups that fold like a project: Images, Video, Voice, Voice design and Library, and Skills and Connectors on one Integrations page

Improved

4 changes
  • Before you choose Generate, a media request shows how long it usually takes beside its price, when a time estimate is available
  • A video or voice request holds no more than your per-reply cap, as a reply does
  • Text to speech offers Delivery tags as chips, to set how a line is spoken
  • The aspect-ratio control names its sizes, such as Story 9:16

Notifications, scheduled tasks, and editing in the canvas

v0.15.439: Notifications, scheduled tasks, and editing in the canvas

New

4 changes
  • The bell in the sidebar keeps what happened while you were away: a reply finished, a question waiting for your answer, a reply that stopped, a goal met or paused, a scheduled run, and replies and mentions in Community. Open one to go straight there, or mark them all read
  • Scheduled tasks run a prompt for you in its own chat: once, every day, on weekdays or on the days you choose, at your time. Each run is charged to your monthly allowance up to the task's cap, from $0.05 to $4 a run, and the bell tells you how it went
  • Scheduled in the sidebar lists your tasks and when each runs next. Run one now, pause or resume it, see its past runs and what they cost, edit it or delete it
  • In the canvas, choose lines of a file and say what should change. The form shows the reply's price before you send, the change becomes the file's next version, and Compare shows exactly what changed

Automatic search, chat by chat

v0.15.437: Automatic search, chat by chat

New

1 change
  • Search automatically in a chat's + menu turns automatic web search off or on for that chat alone. It shows while Search the web automatically is on in Settings, and rests while Web search, Research or Code execution is on

Archive chats, and search all your history

v0.15.435: Archive chats, and search all your history

New

1 change
  • Archive a chat from its menu, or several at once with Select on the Workspace page. Archived chats leave the sidebar and your projects, keep everything in them, and open as before. Archived on the Workspace page lists them, to open or unarchive

Improved

1 change
  • Search chats looks through every message you have, not only your latest 1,000, so it finds older chats too

Fixed

1 change
  • Follow-up suggestions under replies and automatic chat titles work again

Import your chats, and next-message predictions

v0.15.434: Import your chats, and next-message predictions

New

5 changes
  • Import your chats from ChatGPT or Claude in Settings → Account: choose the export you downloaded, and each imported reply says where it was written. Text comes across; pictures, files and code runs stay in your export. A large export is read without freezing the page
  • Search the web automatically (Settings → AI behavior): a reply searches the web on its own when your message needs current information. It’s off until you turn it on, on each device, and a reply that searches is charged like any Web search
  • Charts in replies: bar, line, area, scatter and pie charts drawn in your browser, each with its table and a PNG or CSV to download
  • Save a reply as a Word document or a PDF, and a table or a chart’s data as an Excel workbook. They’re made in your browser, and nothing is sent or kept
  • After a reply, Predictions suggest the message you might send next, in your own words. Tab puts it in the message box. Suggestions as you type are gone

Improved

1 change
  • Dark’s message box placeholder and Light’s count of added lines in Files are easier to read

Fixed

1 change
  • Search on a model without its own search tools works again for members who have set a display name

Export your chats, an app to install and Community outlines

v0.15.433: Export your chats, an app to install and Community outlines

New

6 changes
  • Export every chat in one ZIP from Settings → Account, each as Markdown and JSON. The file is built in your browser and nothing is kept on our servers. Project files, instructions and memory notes aren’t included
  • Install the web chat as an app from your browser
  • Select several chats on the Workspace page to move them to a project or to Trash at once
  • Feature requests and bug reports in Community start with an outline to fill in, and can be up to 5,000 characters
  • Posts and comments from the Entevra team carry an “Entevra team” badge
  • Search chats (⌘K, or Ctrl+K on Windows) can turn Web search, Research and Memory on or off, and open Compact, Share, New project, Skills and Trash

Fixed

2 changes
  • Streaming replies no longer drop letters, often a word’s final “t”, while they arrive
  • A project’s page no longer shows “Saved.” beside a change you haven’t saved yet

Security

2 changes
  • A skill’s name can no longer use lookalike letters to pass for Entevra
  • Shared links never show an author name that passes for Entevra

Steer joins a running reply, and longer sign-ins

v0.15.432: Steer joins a running reply, and longer sign-ins

Improved

3 changes
  • Steer now adds your message to a reply that’s searching the web, researching, or using your memory or a project. The reply takes it at its next step and keeps going, instead of stopping
  • Choose Queue or Steer in Settings → While a reply is running. ⌘Enter, or Ctrl+Enter on Windows, sends a single message the other way
  • You stay signed in for 30 days after you last used Entevra, and for up to 90 days in all before you sign in again

Security

1 change
  • A display name can no longer be made to look like it belongs to Entevra

A clearer Projects sidebar and wider goal budgets

v0.15.431: A clearer Projects sidebar and wider goal budgets

Improved

3 changes
  • The sidebar’s Projects section has its own icon, and a line sets the Entevra Workspace apart from your projects
  • Goal budgets go from $1 to $100, or any amount up to $10,000, or No goal limit. Your account’s own limits still apply
  • On entevra.com, the home animation’s dots gather straight into a small logo above the controls, and Helix plays first for new visitors

Ten starter prompts and a clearer goal budget

v0.15.430: Ten starter prompts and a clearer goal budget

Improved

4 changes
  • The home shows ten starter prompts, led by a system diagram, a landing page, an arcade game and a data dashboard
  • A goal shows what it has spent, and while a step runs, the most that step can still cost beside it, never counted as spent
  • On entevra.com, the home animation starts with Ink, then Helix and Event Horizon, and the logo it draws settles at the mark’s size, evenly spaced above the controls
  • Community limits how often each account can post, comment, edit and like, to keep out spam

Entevra Workspace, custom instructions and predictions

v0.15.429: Entevra Workspace, custom instructions and predictions

New

6 changes
  • The Entevra Workspace comes first among your projects and holds every chat that isn’t in one
  • Pin projects to the top, and sort them by recent activity, most used, name or when they were created
  • Pinned chats are marked inside each project, and Move to… takes a chat into a project or back to the Workspace
  • In a project’s chat, a reply can search and read the project’s other chats when it needs them. Turn it off in Settings → Personalization
  • Custom instructions in Settings → Personalization: tell Entevra about yourself and how it should respond. They’re sent with each new message, except with Web search or Research
  • Predictions as you type: Entevra suggests how to finish your message, and Tab takes it. They’re free, and you can turn them off in Settings

Improved

3 changes
  • Choose a default budget for /goal in Settings, from $1 to $25 ($5 unless you change it), and change the budget of a goal that’s running
  • A goal’s check says which item is still open, and why
  • On entevra.com, every home animation ends on the Entevra mark, which comes alive while you’re idle, as it does in the chat

Projects, editing any message, and goals

v0.15.428: Projects, editing any message, and goals

New

7 changes
  • Projects keep related chats together. Give a project instructions and files, and every chat in it works with them
  • Each project has its own memory notes, which only its chats can read and change
  • Projects sit in the sidebar as folders, each with an icon or a picture you choose
  • Edit any earlier message, not only the last one. The conversation continues from your edit, and the earlier version is a click away
  • Branch in new chat starts a new conversation from any message, with everything up to it. Nothing is charged until you send
  • Send while a reply is still running: your message waits and goes when the reply finishes. Choose Steer in Settings to stop the reply where it is and send yours next
  • /goal: describe an objective, and Entevra plans the steps, works through them and checks each one against what it finds, within the goal’s budget

Improved

8 changes
  • A reply finishes and is kept even if you close the tab or lose your connection
  • File edits in a reply show how many lines were added and removed, and open to the changes
  • A reply’s steps are listed in the order it worked: thinking, searching, then thinking again
  • A file path in a reply opens that file in Files, at the line it names
  • Long conversations scroll and stream more smoothly
  • Ask Entevra to remember or forget something and it updates your memory notes, or says plainly why it couldn’t
  • A message too long to send is stopped before it uploads, stays in the message box and shows its size
  • The model list stays available even when part of it can’t be refreshed

Fixed

3 changes
  • An unusual character in a message can no longer stop a chat from opening
  • A reply saved from two open tabs appears once
  • Chat and project menus in the phone sidebar open in full

Security

1 change
  • Updated build tools to close reported vulnerabilities

Long chats, chapters and smoother replies

v0.15.427: Long chats, chapters and smoother replies

New

5 changes
  • Compact summarises the earlier part of a long chat so it can keep going. It’s in the chat’s usage panel, or type /compact, and it shows its price first
  • Chapters: day dividers, a rail beside the conversation for jumping between its parts, and a list of every chapter
  • 200 starter prompts, with search
  • Turn on notifications to hear when a reply finishes, needs your answer or stops with an error; the tab’s title shows it too
  • Search the skills library even with a typo, sort it, and browse Featured skills

Improved

8 changes
  • Replies appear at an even pace. Tables grow a row at a time, and math and diagrams draw while the reply is still arriving
  • The view follows a streaming reply smoothly, and holds still while you select text
  • Thinking is one row with its time, and keeps that time in your history, on other devices and on shared links
  • Follow-up suggestions arrive one by one. Pick one with the arrow keys, or Shift-click to edit it before sending
  • Generated pictures appear in the chat as soon as they’re ready
  • Images, Video and Voice open with a model already chosen
  • Times follow your device’s 12- or 24-hour clock
  • Global Radio has more than 5,000 checked stations from 204 countries

Fixed

1 change
  • Phones no longer scroll sideways after a long thought

Voice input, a new look and a status page

v0.15.426: Voice input, a new look and a status page

New

3 changes
  • Dictate a message in any browser. The language you speak is detected, and it uses your monthly budget like the writing tools
  • See how each part of Entevra is running at status.entevra.com
  • Avatars: emoji and illustrations, on a colour you choose

Improved

6 changes
  • A new look across the site and the chat: Light and Dark themes, our own typeface, Entevra Sans, and more compact controls
  • A new home page at entevra.com, with a choice of animations
  • Entevra IDE’s sign-in page in your browser is now the same as the web sign-in
  • Sign-in is better protected against automated sign-ups
  • A short notice explains the cookies that keep you signed in
  • A link to a missing page now offers the nearest page instead

Follow-ups and Entevra IDE updates

v0.15.425: Follow-ups and Entevra IDE updates

Improved

3 changes
  • Follow-up suggestions are written as your next message, ready to send
  • Appearance shows each theme as a card, with System as its own choice
  • Pages show the Entevra mark while they load

New

3 changes
  • Older Entevra IDE builds ask you to update before they connect
  • Signing in to Entevra IDE on one computer more than your account allows signs out the one you used least recently
  • When Entevra IDE can’t be allowed for your account, the page says why

Skills, Community images and account deletion

v0.15.424: Skills, Community images and account deletion

New

6 changes
  • Skills: pick one with / in the message box or from the library, write your own, and Entevra suggests one when it fits
  • Community posts and comments can carry an image, checked for safety before anyone sees it
  • Regenerate a reply, switch between its versions and compare two answers side by side
  • Entevra can ask you a question in the middle of a reply and wait for your answer
  • Voice design in the Media studio: describe a voice and hear it speak. The Voice tab is now Text to speech
  • Ask to delete your account from Settings → Account. Until the date shown, you can still read and export your conversations

Improved

5 changes
  • You aren’t charged for a reply that our own restart stops before any of its answer arrives
  • Follow-up suggestions under replies, with a switch in Settings
  • Choose which models you see in the picker, and open Settings as a full page if you prefer
  • Delete a chat permanently, or let Trash empty itself after 30 days
  • Buttons you can’t use yet say why when you point at them

Fixed

3 changes
  • Saving a note to memory works again
  • A ZIP with an empty file attaches, leaving the empty file out
  • Community images open full size, and several load at once

Entevra IDE sign-in and shared links

v0.15.423: Entevra IDE sign-in and shared links

New

3 changes
  • Entevra IDE signs in through your browser. Settings → Account lists the computers signed in and can sign them all out
  • Settings → Sharing lists every shared link to open, copy or revoke
  • Generated images open full size, in a conversation and on a shared link

Improved

3 changes
  • Rewritten Privacy Policy, Terms of Service and documentation
  • Clearer messages when a reply stops early or isn’t saved
  • Usage history, Files and Memory fit a phone screen

Fixed

2 changes
  • The writing tools’ notice clears once you edit the revised draft
  • Chat search shows how many chats match

Entevra, with images, video and voice in your chats

v0.15.422: Entevra, with images, video and voice in your chats

New

5 changes
  • mfdl.ai is now Entevra, at entevra.com
  • Create images, video and voice in any conversation. The results stay in the chat
  • Shared conversations show each reply with its model, its reasoning and any media it made
  • Change your display name and picture, or choose a ready-made avatar
  • Like comments in Community

Improved

3 changes
  • Every page, menu and setting redesigned in one compact style, in Light, Paper and Dark
  • Attachments are ready at once, and PDF, Word, Excel and ZIP files are converted in your browser
  • Long replies keep streaming until they finish
Show earlier platform updates (mfdl.ai)Hide earlier platform updates (mfdl.ai)59

These updates are from mfdl.ai, the platform Entevra replaced on 9 Oct 2026. They’re kept as they were written.

New flagships

v0.15.418mfdl.ai: New flagships

New

4 changes
  • Claude Opus 5.5 and GPT-6 Sol join the main model list
  • Grok 4.7 takes the xAI seat; GPT-6 Luna, MiMo V2.6 and DeepSeek V4.1 Flash arrive
  • Command A+, Fugu Max and Ultra v2, Ember-1, Qwen3.8 Omni Flash, Ling VL and Solar Mini 4
  • Space Bunny Alpha joins the free models

Fixed

2 changes
  • Mercury 2.5 runs on the released model instead of the withdrawn preview
  • Conversations started on Opus 4.8 Fast keep working on Opus 4.8

Astra

v0.15.415mfdl.ai: Astra

New

2 changes
  • GPT-6 Astra and Astra Pro join the model list; Astra takes the OpenAI seat in the picker
  • Ling 3.0 Flash Sante, a health-focused model, joins the free tier

Fixed

2 changes
  • Qwen3.8 Max keeps working after its provider renamed it
  • Two free MiniMax models retired by their provider stay visible in older conversations

A page for every release

v0.15.413mfdl.ai: A page for every release

New

2 changes
  • Every release now has its own page, for example /changelog/v15-413
  • The home page describes the app properly to visitors who are not signed in

Improved

2 changes
  • Privacy and Terms live at one address; the old links land on the right section
  • Docs pages carry their own titles, headings and share previews

Fable 5.1 and a quieter memory

v0.15.411mfdl.ai: Fable 5.1 and a quieter memory

Models

3 changes
  • Added Claude Fable 5.1, now first in the picker
  • Added Muse Spark 1.3, Gemini 3.8 Flash, Mercury 2.5, Hy4 Preview and Granite 4.2
  • KAT-Coder Air was retired by its provider; older conversations still show it

Improved

2 changes
  • Memory now works quietly instead of announcing itself
  • Usage now shows the exact cost of each reply

Fixed

1 change
  • Conversations with memory and web search on used far more tokens than they needed to

Effort on every model

v0.15.407mfdl.ai: Effort on every model

Improved

2 changes
  • Reasoning effort can now be set on almost every model, not just Claude
  • Code execution is available on Claude Haiku 4.5

Fixed

1 change
  • Your effort choice was ignored outside Claude and every reply ran at high

Fairer limits

v0.15.401mfdl.ai: Fairer limits

Improved

2 changes
  • Free models now have fair-use limits
  • Shared pages open faster and large attachments start arriving sooner

Fixed

2 changes
  • Fixed attachments on shared pages showing placeholder text
  • Fixed image generation stopping early on longer batches

Security

1 change
  • Routine hardening across sign-in and shared links

Flagships up front

v0.15.398mfdl.ai: Flagships up front

Models

2 changes
  • Model picker now leads with Claude Fable 5, Grok 4.6, GLM 5.3, Gemini 3.7 Flash and Qwen3.8-Max
  • Previous versions remain under More models; a saved older model stays selected

Code execution and model updates

v0.15.397mfdl.ai: Code execution and model updates

Improved

1 change
  • Code execution runs in a sandboxed Linux shell; Python is available, internet is not

Models

4 changes
  • Claude Fable 5 is now unlocked and selectable
  • Devstral 2 is back online at its provider
  • Added MiniMax M3 and M2.7 (free; M3 with 1M context), Ling 3.0 Flash Fin (free), GLM 5.3 Flash and Qwen3.8 Flash
  • Ox Alpha and GPT-OSS 20B were retired by their providers; older conversations still show them

Fixed

1 change
  • Code execution works again

Also

1 change
  • Files created by code runs are temporarily not attached; printed output still shows in full

Eleven more models

v0.15.396mfdl.ai: Eleven more models

Models

3 changes
  • Added free models Ox Alpha (1M context, vision, tools), Dots3 Note Preview (vision, 512K) and LFM2.5
  • Added Muse Spark 1.2 Contributor, Muse Glimmer 30B, DeepSeek V4 Flash Vision and Solar Pro 4
  • Added Qwen3.8 27B, the Qwen3.8 flagship, ByteDance Seed 2.1 Turbo and Seed 2.0 Code

Fixed

1 change
  • Enabled web search on Gemini 3.7 Flash, Grok 4.6, Sakana Namazu and Qwen3.8 Max

Dark follows your system

v0.15.394mfdl.ai: Dark follows your system

New

1 change
  • The app follows your system dark mode until you pick a theme yourself

Models

1 change
  • Added Grok 4.6, Gemini 3.7 Flash, GLM 5.3, Qwen3.8 Max and Sakana Namazu; Muse Spark and DeepSeek V4 Pro moved to latest releases

Fixed

2 changes
  • DeepSeek V4 Flash and V4 Pro are charged at their listed prices
  • Devstral 2 was retired by its provider; older conversations still show it

Reliability and accessibility

v0.15.389mfdl.ai: Reliability and accessibility

New

1 change
  • Installable as an app from a phone home screen or the browser

Fixed

4 changes
  • Fixed duplicate sends from a fast double Enter
  • Fixed files from one chat overwriting another when switching quickly
  • Fixed Stop corrupting the saved partial reply
  • Fixed the composer sitting behind the browser toolbar on phones

Improved

2 changes
  • Exported conversations name the model behind every reply
  • Improved keyboard and screen-reader support across chats, dialogs and the composer

Stone theme

v0.15.388mfdl.ai: Stone theme

New

1 change
  • New Stone theme replaces Glacier; Glacier users are moved over automatically

Improved

1 change
  • Restyled the See what’s new button; the changelog back link returns to the page you came from

Documentation

v0.15.381mfdl.ai: Documentation

New

1 change
  • Nine documentation pages are now available, linked from the sidebar and account menus

Improved

1 change
  • Tightened the Settings panes and corrected an outdated theme count

Gray default theme

v0.15.380mfdl.ai: Gray default theme

New

1 change
  • Gray is now the default theme; a theme you pick yourself still sticks

Improved

1 change
  • See what’s new button on Gray now matches Paper’s quieter ink style

Fixed

1 change
  • Fixed a white flash on page load in the dark theme

Sitewide announcement strip

v0.15.379mfdl.ai: Sitewide announcement strip

New

3 changes
  • Announcement strip now appears on every page, themed to match each palette
  • Privacy, Terms and changelog links moved into the strip, replacing the footer and home cards
  • Later hides the strip for this visit; Never keeps it hidden until the next release

Improved

2 changes
  • Home particle field is off by default; enable it from the panel rail or Settings
  • Shared conversation pages no longer show the strip

Fixed

1 change
  • Fixed touch targets for the strip’s controls on phones and tablets

Theme lineup and contrast

v0.15.371mfdl.ai: Theme lineup and contrast

Improved

3 changes
  • Theme family settles at four: Paper, Gray, Glacier and Slate
  • Paper’s original colours are back
  • Better contrast in all four themes

Fixed

1 change
  • Fixed unreadable hover states on the strip’s LATER and NEVER buttons in light themes

Answer formatting

v0.15.347mfdl.ai: Answer formatting

Improved

1 change
  • Answers now render with structured headings, links, inline code and quotes

Fixed

2 changes
  • Bold text no longer appears darker than headings
  • Fixed headings falling back to body style on shared and read-only pages

Accurate usage figures

v0.15.344mfdl.ai: Accurate usage figures

Fixed

1 change
  • Fixed the Messages figure counting billable actions instead of messages sent

Improved

2 changes
  • The Messages figure counts only messages you sent
  • Each usage figure explains what it counts

Also

1 change
  • Token and spend figures are unchanged

Announcement banner and default theme

v0.15.340mfdl.ai: Announcement banner and default theme

New

2 changes
  • Announcement banner at the top of the app, dismissible per visit or permanently in Settings
  • Paper is now the default theme; existing accounts move over once

Improved

1 change
  • Banner renders consistently on every theme; on phones it truncates and dismiss opens on tap

Fixed

1 change
  • Fixed banner height and a dismiss control shown when signed out

Particle controls

v0.15.323mfdl.ai: Particle controls

New

1 change
  • Home particles can be turned off, from the home page or in Settings

Improved

1 change
  • Restyled home particles to a finer, calmer look with distinct button icons

Home page particle field

v0.15.319mfdl.ai: Home page particle field

New

2 changes
  • Interactive particle field on the home page, responding to mouse movement and clicks
  • Glass message box with a placeholder drawn from the particle field on hover

Improved

3 changes
  • Home animation pauses when the tab is hidden or while typing
  • Deepened six themes: Sora, Glacier, Warm, Paper, Night and Ink
  • Simplified the docs shortcuts on the home page to two rows

Fixed

1 change
  • Fixed the model list running off screen and faint stars on night themes

Translate and settings redesign

v0.15.298mfdl.ai: Translate and settings redesign

New

4 changes
  • Translate drafts to English, Arabic, Spanish, French or Italian from the composer
  • A + menu on the composer gathers starter prompts, Improve prompt, grammar and translate tools
  • Copy button on your own messages
  • New themes: Paper, Ink, Mist and Sora

Improved

2 changes
  • Settings redesigned on desktop; phones get full per-section settings
  • Fix grammar sticks to correcting; grammar and Translate run on a faster model

Fixed

1 change
  • Fixed mid-stream chat deletion and low-contrast colours on dimmer themes

Models

2 changes
  • Kimi K3 is the default for new chats, with web search on from the start
  • DeepSeek V4 Flash added to the main model list

Also

1 change
  • Retired the Sky, Sage, Void, Hush and Dark themes; existing users migrate automatically

Theme readability

v0.15.284mfdl.ai: Theme readability

Improved

2 changes
  • Labels, toggles and gradient text retuned for readability on Glacier and Hush
  • Focus outlines, text selection and scrollbars now follow the active theme

Fixed

3 changes
  • Fixed unreadable code in research traces on Night and Void
  • Fixed faint error messages and dim secondary text on the dark themes
  • Dialogs on Night and Void darken the page behind them again

Every theme redrawn

v0.15.283mfdl.ai: Every theme redrawn

Improved

3 changes
  • All nine themes rebuilt so each is visually distinct and purpose-built
  • Glacier is now deep violet; Hush is dimmer and cooler
  • Theme buttons describe each theme on hover

Fixed

2 changes
  • Fixed faint delete and warning colours on the Dark theme
  • File drop zone now has its own background in every theme

New themes and instant loads

v0.15.282mfdl.ai: New themes and instant loads

New

2 changes
  • Two light themes, Sage and Hush
  • Privacy Policy and Terms rewritten in plain language on a single page

Improved

4 changes
  • Chats, shared conversations and community threads now open instantly
  • Code execution keeps its sandbox for the whole conversation; files and packages persist
  • Redesigned Changelog, Privacy and Terms, and Community around a pinned contents index
  • Clearer headings and more readable reply text

Fixed

2 changes
  • Fixed code execution failing when web search ran in the same turn
  • Replies stay consistent with earlier work in the same conversation

Also

1 change
  • Collapsed sidebar adds quick links to Settings, Community, Changelog and legal pages

Structured reasoning traces, stream recovery

v0.15.246mfdl.ai: Structured reasoning traces, stream recovery

New

1 change
  • Reasoning traces now stream in titled sections and collapse into an expandable outline

Improved

4 changes
  • Improved syntax highlighting and a redesigned artifact card with live write status
  • Dropped connections now retry automatically
  • Retuned code execution and web search across frontier models
  • Cleaner sign-in with inline errors and code entry

Fixed

3 changes
  • Stopped responses keep all streamed content, and Continue resumes from it
  • Extended thinking now displays on Claude Opus 5
  • The free Poolside coding model works again, including in saved conversations

Models

3 changes
  • Added Qwen3.7 Flash (1M context, low cost)
  • Added GPT-5.6 Terra Pro and Sol Pro alongside Luna Pro
  • Gemini 3.5 Flash Lite replaces the 3.1 tier

Voice input, seven themes, 100 templates

v0.15.198mfdl.ai: Voice input, seven themes, 100 templates

New

5 changes
  • Voice input in English and Arabic (Egyptian, Syrian and Gulf variants)
  • Transcripts are cleaned up for review before sending; nothing sends automatically
  • Grammar button fixes spelling, punctuation and grammar without changing meaning
  • Seven themes, switchable from Settings, the artifact panel and shared chats
  • Mic has a language switch; Ctrl+Shift+M starts listening and Esc stops

Improved

4 changes
  • Template library rewritten and expanded from 43 to 100
  • Better contrast on every theme
  • Redrawn sidebar with grouped navigation
  • Voice capture no longer drops the last phrase

Models

3 changes
  • Kimi K3 moved to the top of the model picker
  • Every image model now edits attachments
  • Image batch generation in sets of 1, 2, 4 or 10

Improve Prompt, six new models

v0.15.173mfdl.ai: Improve Prompt, six new models

New

1 change
  • Improve Prompt wand rewrites a rough prompt in any language and returns it for editing

Improved

3 changes
  • Redrawn composer and a larger, more readable reasoning trace
  • Sending scrolls straight to the reply
  • Templates reorganized into scrollable categories, opened from the lightbulb

Fixed

3 changes
  • Long conversations no longer drop replies on large contexts
  • Code view wraps long lines
  • Usage meter shows the exact enforced limit

Models

3 changes
  • Added Inkling, Thinking Machines’ debut model
  • Added Kimi K3, GPT-5.6 Sol, GLM 5.2, MiniMax M3 and KAT-Coder Air
  • Free tier refreshed with Tencent Hy3, Nemotron 3 Ultra (1M context) and GPT-OSS 20B; retired routes replaced

25 models, image editing, Fusion

mfdl.ai: 25 models, image editing, Fusion

New

4 changes
  • Edit generated images by asking in chat
  • Fusion asks several models at once and blends the strongest parts of their answers
  • Global Radio: live stations on a spinning globe
  • One-tap starter prompts on home, each pinned to its best model

Improved

1 change
  • Faster model switching and a tidier mobile chat layout

Models

1 change
  • Added 25 models across reasoning, coding, vision and web research, several free

Also

1 change
  • Assorted small fixes throughout

Per-chat links, reliability fixes

v0.15.127mfdl.ai: Per-chat links, reliability fixes

New

1 change
  • Every conversation has its own URL that can be bookmarked and shared

Fixed

6 changes
  • Replies that hit the length limit say so and can continue
  • Memory works reliably across Claude models
  • Stopping a response cancels it cleanly
  • Image and PDF uploads reach more models
  • Fixed code-viewer stalls on large files and wrong model labels on shared conversations
  • Steadier mobile layout and quieter theming on legal and changelog pages

Security

1 change
  • Security and reliability hardening across the app

GPT-5.6, Grok 4.5, doubled limit

v0.15.111mfdl.ai: GPT-5.6, Grok 4.5, doubled limit

Improved

4 changes
  • Monthly allowance doubled
  • Web search and extended thinking are off by default for new chats
  • Usage updates instantly
  • Code viewer and phone files-panel refinements

Fixed

1 change
  • Fixed web search on models where it had stopped working

Models

3 changes
  • Added the GPT-5.6 family: Sol, Terra, Luna
  • Added Grok 4.5, Gemini 3.1 Pro, Mistral Medium 3.5, Hunyuan 3, Laguna XS 2.1 and free North Mini Code
  • Free models available after the monthly cap

Stats, handoffs, file viewer

mfdl.ai: Stats, handoffs, file viewer

New

2 changes
  • Summarize any conversation into a shareable handoff brief
  • Platform statistics on the home card

Improved

2 changes
  • Usage and platform stats load instantly
  • Files panel redesigned: line numbers, markdown previews and large-file support

Sonnet 4.6, image generation, page previews

mfdl.ai: Sonnet 4.6, image generation, page previews

New

2 changes
  • Generate images inside a chat
  • Web pages built by models appear as previewable files, openable in the browser

Improved

3 changes
  • Attachments are saved with the chat and included when shared
  • Conversation files easier to find in the files panel
  • Dark-mode and sign-in refinements

Models

3 changes
  • Added Claude Sonnet 4.6
  • Wider model lineup, several free
  • Web search on more models, with sources in answers

Also

1 change
  • Reliability and polish fixes throughout

HTML preview and model picker

mfdl.ai: HTML preview and model picker

Improved

6 changes
  • HTML preview matches real browser rendering, including interactivity and animations
  • Artifacts panel on all screen sizes and shared chat links; auto-opens on phones during streaming
  • Model dropdown rebuilt: Opus 4.7, Sonnet 4.5, Gemini 3.1 Pro, GPT-5.5 and Grok 4.3, with more in a flyout
  • Research mode moved into the model dropdown
  • Empty chat screen simplified to just the input bar
  • Files being written show a shimmer on their card

Community launch, Qwen 3-Max

mfdl.ai: Community launch, Qwen 3-Max

New

2 changes
  • Community: share threads, comment and follow discussions from the sidebar
  • Skills, attachable knowledge libraries, rolling out soon

Improved

1 change
  • Higher usage allowances on many models

Models

2 changes
  • Added Qwen 3-Max for multilingual reasoning and math
  • Cross-conversation memory now limited to Opus 4.7 and Sonnet 4.5