Highlights
- The sidebar’s chat list is rebuilt, with reworked chat groups, drag-and-drop between them, and pinned chats.
- Organize your projects and threads with labels you create, and filter to them from a new “By Label” view.
- Queued follow-ups are steadier: one that runs automatically now streams its progress live, and a bug that could freeze a whole thread is fixed.
- Confidential mode can now analyze images, alongside the image generation it already supported.
- The sidebar’s chat list is rebuilt from the ground up: chat groups, drag-and-drop between them, projects, and pinned chats. The unread dot shows the right color when the sidebar is collapsed and clears once you delete a thread.
- Label your projects and threads however you like, give a thread more than one label, and switch to a new “By Label” view in the sidebar.
- Cmd/Ctrl+1–9 chat switching now works while your cursor is in the composer or another input, instead of being read as typing.
- Your name in the sidebar updates as soon as you rename yourself in your profile.
- A queued follow-up that starts the moment a turn settles now streams its work log and reply live, instead of sitting on a blank “Working” until it finishes.
- A bug that could freeze a whole thread, so you could not send, stop, or edit the queue, is fixed. Any thread still stuck recovers on its own.
- Longer audio and voice recordings no longer time out when you upload them.
- Fluso can now search across your past conversations as project memory, so it reuses earlier work and recalls what you discussed instead of starting cold.
- The accept and reject feedback you give on tasks now shapes how Fluso handles similar work later.
- The assistant is more consistent about when to ask versus act: before committing to a long run, it asks one focused question when an unresolved choice would change the result, and starts right away when the request is already clear.
- Flow diagrams render more reliably: labels fit their boxes, connectors stay attached to shapes, and dense diagrams no longer fail to draw.
- Image analysis is available in confidential mode chats, so that work stays on the private path.
Highlights
- Messages you send while the assistant is working now queue to run next, instead of steering the reply that’s already in progress. Stop is the one direct way to interrupt a run.
- Non-image attachments show as compact cards, so files no longer crowd out the composer.
- An open file updates itself when the assistant rewrites it, with no reload.
- Export a project to a ZIP again.
- Reference images you attach or link now carry through when the assistant redesigns or iterates on an image.
- PDFs, Word files, spreadsheets, and other non-image attachments show as compact cards in the composer instead of large previews. Images and voice notes are unchanged.
- New Chat starts clean: attachments and voice notes you just sent no longer linger in the dashboard composer when you come back.
- A ZIP you attach uploads as the archive itself, instead of being unpacked into separate files.
- An already-open file viewer picks up a replaced file automatically, so you don’t reload to see the latest version.
- A chat you start from the Tasks page gets a real, generated title instead of staying on “New thread”.
- When a lot is syncing in from your connectors, the app holds up better, with fewer errors and timeouts.
- Queued edits, cancels, and follow-ups no longer hang when the assistant is busy or still warming up, and a retried edit can’t apply twice.
- A queued edit or retry that finishes after you’ve switched chats no longer drops its draft or error into the chat you’re viewing.
- When you attach or link reference images, the assistant keeps them through redesign and follow-up generations, instead of dropping them.
- On sponsored accounts, the billing page shows your Confidential plan clearly instead of a contradictory “Basic” marker, with the full comparison behind a “Compare plans” toggle.
Highlights
- Bring your context in from ChatGPT, Claude, or another agent app (Codex and more) — import turns an old export into real Fluso projects, with the decisions, notes, and files intact, so you’re not starting from zero.
- Projects now have working memory — Fluso keeps a running sense of what’s done, what’s open, and how you like things handled, so picking one back up never feels like starting over.
- Steer a running reply, or queue up what’s next — durable enough to walk away for hours, or a day, and find it exactly as you left it.
- Tasks that use your connected apps (GitHub, Linear, Slack, and more) finish in about half the time, with far fewer wrong-tool guesses.
- A hung or unresponsive assistant now recovers in a couple of minutes with nothing lost, instead of a wait that could previously stretch past half an hour.
- Running a task from a task card now gives the assistant the full task details right away, not just its title.
- In-app voice messages show their real length instead of a stuck 0:00.
- A Markdown file in the file viewer shows one Download button instead of two.
- The assistant is lighter on its feet now: old tool results get cleared out once they’re no longer needed, instead of being carried forward turn after turn, so a task takes fewer tool calls and uses less of its token budget along the way.
- The assistant reaches for the right app-specific guide before it acts on a connector, instead of guessing.
- Send a message while the assistant is working and it either steers the current reply or joins an editable queue to run next, with edit, send-now, and cancel controls. The queue lives on the server, not your browser tab, so you can line up a stretch of work, close the laptop, and come back an hour later, or the next day, to find it exactly where you left it — steer and queue are built for real, unattended runs, not just quick back-and-forth, backed by the reliability work below.
- A dropped connection partway through a long file or command now retries and resumes on its own instead of failing the turn.
- The action chips under a task no longer get permanently stuck if one model hiccups — it falls back automatically, with a Retry button for a genuine network blip.
- Reloading mid-reply, or leaving a tab open a while, can no longer leave a conversation showing your message with no answer, or blank it entirely.
- Stopping a turn now reads “Stopped after Xs” instead of the misleading “Worked for Xs”.
- Working memory means each project remembers itself: open items, recent files, and any convention you’ve taught it (a house rule, a style, a correction) carry forward automatically, on top of your regular preferences, and stay scoped to that project. Come back after a few weeks and ask “where did we leave off” and you’ll get a real answer, not a blank page.
- The assistant reads and creates Word documents more reliably now: no hidden formatting junk when it reads one, and real Word tables, not raw Markdown, in the documents it creates for you.
- The import flow lives on the Projects page: upload the export and Fluso sorts it into projects automatically, with a running thread you can watch while it works and a summary when it’s done.
- An expired sign-in code can now be resent from the verification screen instead of starting over.
- Confidential Mode is now live-verified: the Encrypted badge opens a view that cryptographically re-checks the confidential server in real time, covering the attested CPU and GPU and the model being served. The plan is also renamed to Confidential for sponsored accounts.
- A focused internal security audit closed six findings this cycle, spanning token validation, account isolation, and file-handling safety.
Highlights
- Edit a delivered file right in the chat, and your manual changes are kept.
- Fluso is now powered by GLM 5.2, with a per-chat Fast mode for the turns where you want a quick answer over a deep one.
- Replies come back faster, and multi-step turns finish sooner.
- The app is time-aware: relative times like “tomorrow at 9” resolve to your local timezone.
- Finished replies settle cleanly, with no stranded answers or a stuck “Working”.
- Consistent icons and better dark-mode contrast across the app.
- The green “chat finished” dots in the sidebar survive a reload instead of vanishing.
- Links only trigger on real web addresses, so ordinary dotted text isn’t turned into a link.
- Email stays draft-first: the assistant writes to your Gmail drafts and sends only after you confirm.
- Adding tool servers like Linear from chat now completes sign-in instead of failing partway.
- When you ask about your email, the assistant reads across your recent messages instead of occasionally checking only the latest one.
- The assistant no longer strands a finished reply, lingers on “Working” after it’s done, or replays a previous turn’s answer into a new message.
- Brief provider hiccups mid-turn recover on their own instead of ending the reply.
- The inline editor opens Markdown, documents, and code, with a per-thread list of the files in that chat. Your edits are handed back to the assistant instead of being overwritten the next time it works on the file.
- Fast mode lives in the composer and stays set for that thread. On Telegram, send
/fastto toggle Fast for the current Telegram conversation. - The assistant is clearer about what it can and can’t do, and points you to the docs for how it works rather than claiming features that aren’t set up.
- After you upgrade your plan, the whole app reflects it right away instead of showing the old plan for a few minutes.
- Connected-account actions are locked to your own account.
- File uploads work again while confidential mode is on.
Highlights
- Confidential mode on the Max plan: run your chats and voice through end-to-end encrypted inference, so your work stays on a private path.
- Manage your connected tool servers in one place, with per-tool permissions and a clean reconnect.
- The assistant recovers from mid-task hiccups instead of hanging on “Working”.
- Report a bug from inside the app and it becomes a tracked issue the team follows up on.
- Attachments stay put when you switch chats: image previews and files you haven’t sent yet are kept per conversation.
- Sidebar chat titles start from your first message and refine once your intent is clear, and never overwrite a title you set yourself.
- The bug-report dialog now takes PDF and Markdown files, and you can paste files straight in.
- See your connected MCP servers on one screen, set what each tool is allowed to do, and reconnect without dropping your current connection. Credentials are stored encrypted on the server, not in local files.
- More tool servers connect on the first try and stay connected, so providers like Supabase and Linear no longer quietly expire.
- If you have many connected apps, the ones past the first ten no longer show as “not connected”.
- The assistant recovers from mid-task network and provider hiccups instead of leaving you on “Working”, and stops cleanly when you ask it to.
- Voice notes come back as proper sentences, with punctuation and capitalisation.
- Emails the assistant drafts arrive as properly formatted HTML and sign off with “Sent via fluso.ai”, which you can turn off or rename. You still approve every message before it sends.
- On the Max plan, turn on “Confidential APIs” from settings or the chat composer. It’s a one-way switch: while it’s on, file uploads, image generation, and background syncing pause so nothing leaves the private path.
- A new Beta badge on the home screen sets expectations and gives you a one-click way to report a bug.
- API keys your tools use can no longer be read back through the assistant’s shell.
Highlights
- A cleaner, more cohesive app after a top-to-bottom design pass.
- Search your threads right from the sidebar.
- The assistant got noticeably smarter at research, drafting, and visuals.
- Normalised typography, spacing, and layout across the whole app. Sidebars, headers, buttons, cards, and empty states all feel more consistent.
- A clearer step-details panel in the work log.
- On macOS, the window frame now follows your light/dark theme. No more dark strip at the top on launch.
- Search past conversations from the sidebar, with matches highlighted.
- Telegram chats are now labelled in the sidebar so you can tell sources apart at a glance.
- Deep Research asks a clarifying question when your request is ambiguous instead of guessing, and uses refreshed web-search connectors.
- Draft emails and Slack messages get an automatic polish pass before you approve and send, so they read like a person wrote them.
- Slide decks use richer visuals for pitch and brand decks, while data-heavy decks stay clean.
- Sharper image generation, now powered by GPT-5.4 Image 2.
- Backend services now run in the EU region for European users.
- Fewer noisy banners: the “Technical Issue” banner only appears when something user-facing is genuinely broken.
- Security hardening across the app from a recent external review.
Highlights
- The app updates itself now. New versions install quietly in the background.
- Drop your connection mid-answer and the assistant picks back up, with nothing lost.
- Syncs recover on their own instead of getting stuck.
- Background auto-updates: once you’re on 1.0.8+, new builds install on their own. A sidebar dot shows when one is ready, and Settings has a “Check for updates” button.
- Windows and Linux now use the same clean, frameless window style as macOS.
- Windows builds are properly signed (no more “unknown publisher” warning), and the taskbar shows the Fluso icon.
- Gmail and Slack syncs recover on their own from rate limits and hiccups. If a Gmail connection truly expires, you’re clearly prompted to reconnect.
- Queued syncs honestly show as “queued” instead of pretending to run.
- Google Drive: once a file is uploaded, clicking it again opens it in Drive instead of uploading a duplicate.
- If your connection drops or your laptop sleeps mid-response, your answer is preserved and finishes on its own.
- In-progress tasks are no longer lost if something restarts behind the scenes.
- The “Worked for X seconds” timer is honest now, and empty voice notes no longer break the conversation.
- Free and Sponsored Pro plan limits were raised, with more headroom before you hit a cap.
- The name, role, and preferences you set during onboarding now carry across the whole app.
- Open the app with no network and you get a clean “No internet connection” screen with a Try Again button that auto-recovers when you reconnect.
Highlights
- Clearer sync history and status across your connectors.
- A smoother connectors page and onboarding.
- Sync history shows clear, readable failure reasons when something goes wrong.
- Upload files to Google Drive straight from chat.
- The connectors page handles 100+ integrations without locking up, and stalled OAuth connections can be recovered instead of getting stuck.
- Onboarding, sign-in, and project-switching polish throughout.
- Dismiss tasks you don’t want surfaced, and filter to completed-only.
- Open the app offline and you get a clean “No internet” screen that recovers when you reconnect.