Manus Release Notes

Every Team Can Now Be AI Native

In the last 6 months, we have seen a fundamental change in how our team works at Manus.

Our product managers use Manus to consolidate and prioritize product feedback the team posts in Slack. Our team leaders use Manus to assign tasks and track progress. And when a customer support inquiry comes in, our troubleshooting bot checks our codebase before suggesting potential causes and fixes. With Manus, our team is able to stay lean and move extremely quickly like never before. We joke that sometimes we don’t know if we’re talking to a coworker or Manus in Slack.

Yet we have the privilege of being an AI company. We have some of the best engineers and product thinkers who are working to push the frontier of AI applications and, at the same time, doing the same thing within our teams.

With Manus Team Plan and Enterprise Plan, we hope to bring the same power to everyone out there who wants to build an AI-native company, starting today without needing all the engineering resources and staying in the loop with AI 24/7 the way we do.

Why is Manus Team Plan Powerful

In the last few months we have met with dozens of teams who build their entire companies on Manus and here are the reasons that we hear from them on why Manus has been so powerful for them:

• The best LLMs for each task: As a company, you want a team of experts who can specialize in coding, marketing, and creatives, and other things. Just like you won't staff a whole company with software engineers, Manus picks the best LLM model for each part of the task, making sure the team creates the best results possible.
• Ready-to-go outputs: Our team has spent a lot of effort making sure all the deliverables in Manus, ranging from research reports to PowerPoints to dashboards and multimedia outputs, are polished and ready to go
• Idea to a prototype in minutes: Your team may have lots of ideas they want to experiment with, but companies may not always have the resources to support each one of those. With Manus it becomes easier than ever to build. You can have a web app built, hosted, published, even in an SEO-optimized way, with full analytics within minutes
• Multiply your productivity with sub-agents: Manus allows you to deploy hundreds of sub-agents in one task, making it easier than ever to scale up and down your investment. It’s more efficient than ever to work on large scale projects without sacrificing on quality or depth.

How Manus for Business Works

Manus has been built for collaboration since the start. Our tasks live in the cloud. Task collaboration has always been seamless. But over the last few months, we have worked closely with our top team users to make the experience even better.

• Projects that carry context: Share instructions, files, Skills, and connector settings across related tasks. Scheduled tasks can use the same Project context for recurring work, such as a weekly report.
• Skills the whole team can use: Make a useful method available beyond the person who built it. Administrators can enforce a Skill when a process needs to stay consistent.
• Collaboration with clear boundaries: Teammates can contribute to shared tasks. Team Only access keeps supported tasks, replays, files, and internal sites within the workspace. Free Guests can review shared work and comment without taking a paid seat.
• Shared capacity and visibility: Credits from paid seats are pooled, billed centrally, and tracked in a team usage dashboard. Administrators can add credits as needs change.
• Controls for a growing team: Roles, sharing controls, audit logs, SSO, and SCIM help administrators manage access. Manus prohibits model providers from training on Team customer data and holds SOC 2 Type II, ISO 27001, and ISO 27701 certifications. See the Trust Center for details.

How to get started today

Manus Team Plan is available now. It starts at US$20 per paid seat per month or US$200 per paid seat per year for 4,000 credits and scales with your needs; Guest seats are free. Create a team from your Account page or visit the Manus Team Plan page.

Organizations that need managed accounts, legal hold, SIEM log integration, agent network controls, or per-user usage limits can explore Manus Enterprise. Contact sales to discuss the controls your organization needs.

If you want to learn more about how Manus supports companies around the world, check out here:

• Follow Manus for Business on Linkedin. We will share more use cases, best practices on how companies use Manus to become AI native.
• Request an AI audit for your company. We will spend 60 minutes together and deliver you a free AI implementation plan and tech stack suggestion.
• Join us in Manus Executive Dinner series around the world. We have hosted a few in the last months and learned so much and hope to see you at a future one.

We can’t wait to help transform the next thousands of companies into AI-native companies, and see what you build with Manus!

Manus 2.0 is out. For the past month, I’be been making videos with Manus almost every day.

Launch films for new features, game trailers, news explainers, beat-synced animations, a last minute launch film for Cue that was requested the night before it went live. I even handed Manus the footage from a trip and let it do the editing, and was tempted to set up a scheduled task and have it make a new video every day on its own.

Now that we’ve launched, I’d like to explain what I've learned in the past month while it’s still fresh. I will focus on two things: what Manus can do with video, and why, on top of that, we built Video Editor.

The 52-second film below introduces the new video features. It was made by Manus, too.

A Great Shot Is Not A Video Yet

Video models are already remarkably good. Write one sentence and a few minutes later you have a beautiful shot.

But the more videos I made, the clearer it became how far a great shot is from a video you can actually publish. First, you have to figure out what the video is really about. Then come the questions: How do you get people to stop scrolling in the first two seconds? Where should the joke land? What music fits, and should it dip when someone is talking? Which font should the caption use? Which beat should each cut hit? And on top of all that, everything has to match the brand’s tone. Making a video means hundreds of small decisions, and every one of them depends on context.

The second challenge comes after the first version is generated. There’s always something you want to change: a more upbeat song, a reworded caption, a shot trimmed by half a second. If all you have is a chat box, small tweaks like these often mean regenerating the entire video, and the parts you liked in the previous version may change along with them.

So we set ourselves one question:

How can someone who has never learned video editing make a video they want to share?

Our answer has two halves. The first is Manus makes the whole video for you. The second is Video Editor, which hands the final cut back to you.

Part 1: Manus Makes the Whole Video, Not Just the Shots.

For creators: it searches the web for ideas

Most of my videos start with a single sentence. Once I typed only this: "Search recent news, pick any topic, English, YouTube style, lots of memes (you can search for them), keep it fast."

Manus picked Fat Bear Week at Katmai National Park in Alaska, where people vote for the brown bear that got fattest before hibernation. It read that week's news, wrote a script, remade classic memes like Drake, Distracted Boyfriend and This Is Fine with bears, added real photos of the contestants, and turned it all into a 100-second explainer. Every fact on screen cites its source.

At the end, it warned me that one vote count had appeared only on a fan Facebook page and wasn't officially confirmed, so I should check it before posting.

Search is what gives Manus a feel for the internet. It reads what people are posting, not just the news. It can check Reddit for the memes going around, break down how a trending YouTube video is cut, or take a reference link, pull apart its pacing and shots, and make its own version. I once gave it an Instagram link and said only "recreate this video." About an hour later I had the finished video, and a colleague's first reaction was: how is that app UI so real?

For everyone: it makes every piece, and writes code for the rest

Manus can make everything a video needs on the spot: shots from video models like Seedance 2.5, plus images, music, voiceover and sound effects.

But what's special is the code. Manus has its own computer, enabling it to write code for motion graphics, layouts, charts and map routes, render them, screenshot the result to check it, and keep fixing what it doesn't like. I've watched it draw a grid over a frame in Python to locate pixels, and use FFmpeg to turn 24 fps footage into 120 fps so a 5x slow-motion shot stays smooth. I never taught it that. It worked these tricks out to make the video better.

Generated shots and code-made visuals can share one video. For BEAT RUSH, Manus generated a dancing anime character with Seedance, then built the rhythm-game interface and beat effects in code.

For marketing teams: it already knows your brand

The best example is the launch film for Cue.

At 9 PM the night before Cue launched, I got the brief: we need a launch film tomorrow. Before, that would have been close to impossible.

This time I made it in Manus Studio. It read our meeting notes, Notion, Slack, Figma and code repos, plus the preferences it remembers about me. It helped me sharpen the idea, then drew the storyboard, then made the film.

Cue launch film storyboard.

76 minutes in, Manus proposed three directions and I replied only "A. Make it." Then it drew these 16 panels.

After a few hours of polishing, the film was done before launch. Once @CueAgents posted it, it passed 1.7 million views on X.

What I took from it: the hardest part of a video is usually not the visuals. It's matching the brand's tone and this campaign's exact requirements. That takes countless details and context, and they already sit scattered across your everyday work.

One sentence can make any kind of video

Manus made far more kinds of video this month than I expected.

Left to right, top to bottom: product launch film, game trailer, news explainer, AI UGC ad, music video, short drama, code-only animation, viral physics animation, knowledge explainer

The functionality I'd recommend most for small teams is the AI UGC ad: the TikTok product recommendation that looks casually filmed by a real person, like this ad for a collapsible coffee cup. A small brand with no budget for creators can make a version with Manus first and see how people react.

Most of these videos started from a prompt of one or two sentences.

Part 2: Video Editor gives you the final cut

Manus usually gets the first version 80 to 90 percent there. But the last 10 to 20 percent is often what you care about most: the music a bit lower, a caption reworded, a shot half a second shorter, your own product photo instead of a generated one.

These edits are small, but doing them only through chat is awkward. "Move it half a second earlier" is hard to say precisely, and fixing one thing can shift another. So we built Video Editor. It lives in Manus Studio, the Manus desktop app.

Every layer is its own track

Every ingredient Manus uses, from generated shots and images to captions, motion graphics, voiceover, music and sound effects, sits on its own layer of the Video Editor timeline. You don't get a flattened MP4. You get a video you can take apart and change.

A vlog clip split into eight layers, 30 seconds

For everyone: drag it yourself, or just ask Manus

You can drag, trim, rewrite captions and set volume with your mouse, like in any editor. Or tell Manus "turn the music down" or "tighten this part." Both change the same project. When you hand it back, Manus picks up from your edits instead of starting over.

Edit it yourself, or just ask Manus.

Frames from the launch film above

You don't need editing experience to use Video Editor. You only need to know the feel you want.

For creators: start from your own footage

Because Video Editor runs on your desktop, it can start from the footage already on your computer. Drag videos into the chat, or tell Manus which folder they're in, and it imports them and lays them on the timeline.

From the official 2.0 tutorial: drag three local clips into Manus, say "Make a 15-second seaside film," then trim in Video Editor. 14 seconds

I tried the whole workflow with footage from a trip: 125 camera clips, almost two hours in total, on an external drive. Manus imported everything into the editor, transcribed and sampled frames from every clip, and worked out what happened over the three days.

Then it picked the good moments and arranged them into a story that follows the itinerary. The first cut ran 12:48, longer than planned, so it listed each segment's length and lines, cut more than twenty repetitive or slow spots, and brought it to 10:50 with 195 cuts.

When I asked for some variety-show energy, it made over a hundred text effects, stickers and chapter cards and placed each on a track. It split the music into seven cues by mood and dimmed it whenever someone speaks.

Here is the final timeline. Every caption, text effect and music cue is still on its own track, so you can change anything.

The travel film's final timeline: 325 captions, 112 text effects and stickers, 195 video segments, 7 music cues, 118 sound effects. Drawn from the project data, not a screenshot

Marketing teams can use the same workflow to turn long videos into short ones: a meeting recording or an interview can become several clips ready for social.

Final thoughts

This month made me believe one thing more than anything: AI keeps making it easier to make a video, but whether it's "good" still depends on what you want the video to say and how you want it to look.

AI won't replace human creativity. It multiplies it. Manus can find the idea, write the script and make the first version, but the final cut is always yours.

Some of what I made with Manus this month, 16 seconds, music from the launch film. Notion plays a compressed preview

How to start yours?

Choose Video in Manus and describe the video you want in one sentence, on web or desktop. Video Editor is in Manus Studio, the desktop app for macOS and Windows. On the web, click "Edit in Video Editor" and you'll be guided to download the desktop app and edit on the full timeline.

Manus is launching Manus Flex

Manus is launching Manus Flex, a new module that lets users power Manus with their own API key from a supported inference provider.

Since inception, Manus has built a managed agent product that handles model selection, harness, and agent infrastructure, such as web app hosting and databases. Today, we are opening up the Manus harness and agent infrastructure to work with a wider range of inference providers.

With Flex, users can bring an existing provider relationship into Manus. Their API key connects provider inference to the same projects, tools, execution environments, and agent workflows they already use.

Manus continues to provide the agent system that plans the work, uses tools, operates across execution environments, and turns model output into completed tasks. Flex configuration includes a model and a reasoning-effort setting for the connected provider. Model inference is billed directly by that provider. Other services and infrastructure used as part of a Manus task still consume Manus credits.

The Flex Inference Partner Program expands bring-your-own-key access

The Manus Flex Inference Partner Program expands the number of inference providers that users can connect through Flex. It gives builders a clearer path to providers that fit their existing technical and commercial setup.

We are excited to announce OpenRouter, Fireworks, and Modal as our initial inference partners for Manus Flex, providing a wide range of model selections alongside large language model providers.

“OpenRouter is excited to partner with Manus to expand model choice through Manus Flex. Together, we’re making it easier for users to bring their preferred models into a powerful agent experience.”
— Alex Atallah, Co-Founder and CEO at OpenRouter

“Manus Flex gives users more choice and control over the models powering their products. We’re excited to partner with Manus to bring Fireworks’ high-performance model capabilities to Manus Flex, making it seamless for teams to choose the right model for the job.”
— Pranav Jain, Product Lead at Fireworks

“We're thrilled to partner with Manus on Flex. Connecting Modal's open model ecosystem to Manus gives users more choice in the models powering their agents.”
— Adam Azzam, Member of Product Staff

Connect your API key and start building

Manus Flex lets you bring your own API key into the Manus workspace.

Connect a supported inference provider, configure its primary model and effort level, then use your existing provider account in the projects, automations, and agent workflows you are already building.

Explore Manus.

1. A new foundation for going further

Cascade: The latest iteration of our in-house agent harness

It keeps a project light from the start and brings in specialized capabilities only when the work needs them. The brief, the page, the video, and the automation stay part of one connected project.

Better results. Less waste. In one tested configuration, Cascade used 23.2% fewer tokens, finished tasks in 28.2% less time, and cost 32% less to run than our previous system.

Cloud Computer: When your project needs a place of its own

Some projects outgrow one machine. A multiplayer game needs a server that stays up when your laptop closes. An automation needs somewhere to run around the clock.

Now you can purchase a Cloud Computer: a dedicated environment for the projects that need it, from the server behind the game to a permanent home for an automation.

Automations: From scheduled tasks to event-triggered workflows

Scheduled Tasks used to let Manus start work only at a set time. Now, with Automations, work can also begin when something happens in a connected service.

A new email, a change in ad performance, a calendar event, a Slack message, or a Notion update can all trigger a task. Just tell Manus in a single prompt what to watch and what to do next, and it sets up and runs the workflow.

2. Manus Studio: From finishing tasks to making great work

With Manus 2.0, the desktop app is upgraded to Manus Studio: a shared workspace for people and AI.

Each kind of work needs its own professional tools, but carrying every capability into every task makes the same work slower and more expensive. So Studio brings in only what your project calls for, keeping Manus's context focused on the work at hand. In this release, Studio gets major upgrades for both everyday work and professional creation: documents, spreadsheets, PDFs, and slides, as well as websites, code, games, and video. And for video and games, where the bar is especially high, we're introducing two all-new professional environments: Video Editor and Game Dev.

Video Editor: Make the first cut in minutes. Keep the final cut for yourself.

Say your launch ad goes out tomorrow, and Manus has already put together a first cut. It's close. You'd just like a warmer song, and you'd rather use the product shot you took yourself than the generated one.

Before, changes like that meant writing a new prompt and generating the whole video again. In Studio, Manus gives you the timeline instead. The clips, images, text, motion graphics, and audio it used are all there as separate pieces, so you can swap the song, drop in your own footage, move an animation, and export when you're happy with it.

You can also hand it back. Ask Manus to take another pass, and it works from your edits.

Video Editor is built for short product ads (30 to 60 seconds), AI UGC, data motion graphics, tutorials, vlogs, and motion-led videos. You don't need any editing experience.

Alchemy mode: AI as your creative director

Video Editor also has Alchemy mode. You start from an idea, and Manus takes on the creative direction, using video and code generation together. The jump in quality is surprisingly big: the ideas are more creative, visual details are under tighter control, the motion is more polished, and the pacing of the edit is noticeably better. When quality matters most, Alchemy mode is the one to use. You can still fine-tune the result on the timeline afterward.

Game Dev: Build it. Ship it. Play it together.

Most gamers have a game they've wanted to make for years. AI can already get you a decent version. Game Dev is for making the version you actually had in mind.

It uses video, image, and coding models together, and in Max Ultra mode your game looks great, plays well, and feels right in your hands. Start from a playable template and you're playing your idea in the first session.

In the Edit panel, you can watch the game run, manage assets, edit code, and change movement, art, and scenes by hand. If one villager's hair should be silver instead of brown, select that villager and change it. Everything else stays as it was.

When it's ready, publish it as a website that anyone can play from a link, or bring your friends in.

Multiplayer is where most projects like this stall. It needs a server that's always on and reachable by every player, which means hosting, networking, and upkeep most people never signed up to learn. With Game Dev, you don't have to.

Setting it up takes a few clicks:

Configure → Multiplayer → Select Cloud Computer → Buy one.

Manus handles the rest. It works for racing, head-to-head, competitive, and 3D games. Your friends join, and what they find while playing shapes your next version.

Remote Control and Computer Use: Your computer, one voice command away

A remote desktop app lets you reach your computer from your phone, but you still have to do every click yourself on a tiny screen. Now you can tell Manus what you need instead.

Say you're in a cab and need the deck on your computer at home. Ask Manus to find the latest one in your launch folder and send it to you, and watch your desktop live on your phone while it does. If you'd rather have the PDF, just ask.

This runs on Computer Use. In a connected and authorized session, Manus works in its own visible workspace on your computer, using the files, browser, and apps you've approved. It can test your app while you work on something else, or keep working while you're away, and you stay in control of what it does.

3. Cue: A new app from Manus

Cue is a new standalone app for your personal agents, on your phone and desktop, built on the same infrastructure as Manus. Give it a cue, and it gets to work.

Every agent gets its own identity

In Cue, each agent has its own email, phone number, wallet, and computer, so it can send messages, pay within the budget you set, and see a task through on its own machine. It can also take your calls and leave you a summary in Cue.

Build a team

Put several agents in a group chat with a shared goal, and they hand work to each other: one researches venues for your New York launch, another makes a shortlist, a third drafts the deck. You set the direction and make the final call.

Everyday services, handled by agents

Cue also works with the services around you. At a restaurant, scan the QR code and your agent can order for you or hold your place in line.

Build with Manus, build with us

Think back to your first "wow" moment, when an idea became something real right in front of you.

Manus 2.0 is about everything that comes after it. The song you swap. The product shot you drop in. The villager you give silver hair. The friends who join the game you built. The agent that has the answer waiting when you pick up your phone.

That's where an idea becomes yours.

We started Manus to help people turn ideas into something real, and Manus 2.0 keeps that promise.

Manus empowers you to build. You decide what's worth building.

Availability

Manus 2.0 is available now on the web, desktop, and mobile.

Cue is available on the web, desktop, and mobile, with iOS coming soon after App Store review.

Cue is in early access and free to use with an invite code. You can start with MEETCUE, available to a limited number of early users on a first-come, first-served basis. Everyone who gets access also receives extra invite codes to share.

Today, we are thrilled to introduce Manus 2.0.

Manus 2.0 is not a version update. It's a new architecture, new products, and new capabilities built around a fundamental question: what should the next generation of Manus look like?

1. A new foundation for going further

Cascade: The latest iteration of our in-house agent harness

It keeps a project light from the start and brings in specialized capabilities only when the work needs them. The brief, the page, the video, and the automation stay part of one connected project.

Better results. Less waste. In one tested configuration, Cascade used 23.2% fewer tokens, finished tasks in 28.2% less time, and cost 32% less to run than our previous system.

Cloud Computer: When your project needs a place of its own

Some projects outgrow one machine. A multiplayer game needs a server that stays up when your laptop closes. An automation needs somewhere to run around the clock.

Now you can purchase a Cloud Computer: a dedicated environment for the projects that need it, from the server behind the game to a permanent home for an automation.

Automations: From scheduled tasks to event-triggered workflows

Scheduled Tasks used to let Manus start work only at a set time. Now, with Automations, work can also begin when something happens in a connected service.

A new email, a change in ad performance, a calendar event, a Slack message, or a Notion update can all trigger a task. Just tell Manus in a single prompt what to watch and what to do next, and it sets up and runs the workflow.

2. Manus Studio: From finishing tasks to making great work

With Manus 2.0, the desktop app is upgraded to Manus Studio: a shared workspace for people and AI.

Each kind of work needs its own professional tools, but carrying every capability into every task makes the same work slower and more expensive. So Studio brings in only what your project calls for, keeping Manus's context focused on the work at hand. In this release, Studio gets major upgrades for both everyday work and professional creation: documents, spreadsheets, PDFs, and slides, as well as websites, code, games, and video. And for video and games, where the bar is especially high, we're introducing two all-new professional environments: Video Editor and Game Dev.

Video Editor: Make the first cut in minutes. Keep the final cut for yourself.

Say your launch ad goes out tomorrow, and Manus has already put together a first cut. It's close. You'd just like a warmer song, and you'd rather use the product shot you took yourself than the generated one.

Before, changes like that meant writing a new prompt and generating the whole video again. In Studio, Manus gives you the timeline instead. The clips, images, text, motion graphics, and audio it used are all there as separate pieces, so you can swap the song, drop in your own footage, move an animation, and export when you're happy with it.

You can also hand it back. Ask Manus to take another pass, and it works from your edits.

Video Editor is built for short product ads (30 to 60 seconds), AI UGC, data motion graphics, tutorials, vlogs, and motion-led videos. You don't need any editing experience.

Alchemy mode: AI as your creative director

Video Editor also has Alchemy mode. You start from an idea, and Manus takes on the creative direction, using video and code generation together. The jump in quality is surprisingly big: the ideas are more creative, visual details are under tighter control, the motion is more polished, and the pacing of the edit is noticeably better. When quality matters most, Alchemy mode is the one to use. You can still fine-tune the result on the timeline afterward.

Game Dev: Build it. Ship it. Play it together.

Most gamers have a game they've wanted to make for years. AI can already get you a decent version. Game Dev is for making the version you actually had in mind.

It uses video, image, and coding models together, and in Max Ultra mode your game looks great, plays well, and feels right in your hands. Start from a playable template and you're playing your idea in the first session.

In the Edit panel, you can watch the game run, manage assets, edit code, and change movement, art, and scenes by hand. If one villager's hair should be silver instead of brown, select that villager and change it. Everything else stays as it was.

When it's ready, publish it as a website that anyone can play from a link, or bring your friends in.

Multiplayer is where most projects like this stall. It needs a server that's always on and reachable by every player, which means hosting, networking, and upkeep most people never signed up to learn. With Game Dev, you don't have to.

Setting it up takes a few clicks: Configure → Multiplayer → Select Cloud Computer → Buy one.

Manus handles the rest. It works for racing, head-to-head, competitive, and 3D games. Your friends join, and what they find while playing shapes your next version.

Remote Control and Computer Use: Your computer, one voice command away

A remote desktop app lets you reach your computer from your phone, but you still have to do every click yourself on a tiny screen. Now you can tell Manus what you need instead.

Say you're in a cab and need the deck on your computer at home. Ask Manus to find the latest one in your launch folder and send it to you, and watch your desktop live on your phone while it does. If you'd rather have the PDF, just ask.

This runs on Computer Use. In a connected and authorized session, Manus works in its own visible workspace on your computer, using the files, browser, and apps you've approved. It can test your app while you work on something else, or keep working while you're away, and you stay in control of what it does.

3. Cue: A new app from Manus

Cue is a new standalone app for your personal agents, on your phone and desktop, built on the same infrastructure as Manus. Give it a cue, and it gets to work.

Every agent gets its own identity

In Cue, each agent has its own email, phone number, wallet, and computer, so it can send messages, pay within the budget you set, and see a task through on its own machine. It can also take your calls and leave you a summary in Cue.

Build a team

Put several agents in a group chat with a shared goal, and they hand work to each other: one researches venues for your New York launch, another makes a shortlist, a third drafts the deck. You set the direction and make the final call.

Everyday services, handled by agents

Cue also works with the services around you. At a restaurant, scan the QR code and your agent can order for you or hold your place in line.

Build with Manus, build with us

Think back to your first "wow" moment, when an idea became something real right in front of you.

Manus 2.0 is about everything that comes after it. The song you swap. The product shot you drop in. The villager you give silver hair. The friends who join the game you built. The agent that has the answer waiting when you pick up your phone.

That's where an idea becomes yours.

We started Manus to help people turn ideas into something real, and Manus 2.0 keeps that promise.

Manus empowers you to build. You decide what's worth building.

Availability

Manus 2.0 is available now on the web, desktop, and mobile.

Cue is available on the web, desktop, and mobile, with iOS coming soon after App Store review.

Cue is in early access and free to use with an invite code. You can start with MEETCUE, available to a limited number of early users on a first-come, first-served basis. Everyone who gets access also receives extra invite codes to share.

Introducing ElevenLabs on Manus! Generate speech, transcribe recordings, clone voices, and build audio apps by talking to Manus.

Simply describe the audio you need or the app you want to build, and Manus handles the rest. From there, the same chat is where you turn blog posts into podcasts, extract meeting notes, and design sound effects. ElevenLabs keeps doing what it does best in the background — powering world-class AI audio — while Manus handles the day-to-day on top.

The ElevenLabs connector is now live on Manus

Getting Started

The simplest way to begin is to just ask. Tell Manus you want to create some audio — what the script is, whose voice to use, the feel you're going for — and Manus will connect to your authorized ElevenLabs account and start generating. All AI-generated audio should be reviewed before publishing or distribution. You don't have to write code, manage API keys, or set anything up beforehand. By the end of the conversation, you'll have finished audio files, all processed in your own ElevenLabs environment.

When you're ready to start, there's one quick step: connect your account. Click Connect under Integrations, sign in to ElevenLabs, and Manus is ready to generate. Manus will prompt you when it's the right moment, so you don't have to think about it until you need it.

Claiming your ElevenLabs connection from Manus Integrations

What It Can Do for You

If you already use ElevenLabs, tasks that used to require coordinating across multiple tools and scripts now happen in a single chat against the account you already own. This integration bridges your audio workflows directly into your workspace without disrupting anything you already have set up — your ElevenLabs billing and usage history remain governed by your ElevenLabs account terms.

Generate scripts and audio

Manus writes the script and generates the audio before you lift a finger

You describe the content in one prompt. Manus drafts the text, then generates the audio against your connected ElevenLabs account — selecting the right voice, applying the proper pacing, and handing back a finished file.

Use it to:

• Launch a podcast version of your newsletter without recording anything yourself.
• Stand up a voiceover for a product demo video with localized languages.
• Rebuild your onboarding flow around narrated documentation.
• Stage a custom soundscape for a creative project.

Prompts to start with:
"Take my latest blog post, narrate it with a professional British voice, and create a 30-second teaser with background sound effects for Instagram."
"Create a voiceover for this product launch script in Japanese, using a warm and energetic tone."

Manus handling audio generation from the first prompt

Process audio files for insights

Most audio doesn't start as text. It starts as a raw recording. Manus handles that. With the ElevenLabs connector on, upload the audio file securely within your Manus session, and the ElevenLabs connector will seamlessly process it — identifying speakers, drafting summaries, and organizing the insights as it goes.

Use it to:

• Bulk-process customer interviews into clean transcripts and action items.
• Review your meeting recordings for key decisions and organize them into a shared doc.
• Reorganize a rambling voice memo into a structured project brief.
• Extract the five strongest quotes from a podcast episode for social media.

Prompts to start with:
"Here's the raw audio from our all-hands meeting. Transcribe it, label each speaker, and pull the key decisions into a doc."
"Review this customer interview recording. Flag any mentions of pricing or competitors and summarize their main pain points."

Prompting Manus to process audio with the ElevenLabs connector on

Build voice powered applications

With the ElevenLabs connector, Manus can turn your ideas into full voice-powered applications. Ask for a personalized song generator, a language tutor, or an interactive story, and Manus will build the app — using ElevenLabs as the audio engine.

Use it to:

• Clear the hurdle of building a full-stack app by letting Manus write the code.
• Plan a personalized audio experience around a viral trend.
• Double down on a creative idea by giving it a voice.

Prompts to start with:
"Build a personalized song generator web app with a retro UI, support for four languages, and narration using my authorized voice clone. (Note: you must have rights to any voice you clone, and ElevenLabs’ consent requirements apply.)"
"Create a language practice app where users can talk to an AI tutor that responds with natural speech."

Availability

The ElevenLabs connector is available to Manus users with an active ElevenLabs account. Available capabilities depend on your ElevenLabs plan and credit balance. Setup walkthroughs are in the Manus docs at manus.im/docs.

Common Questions

If you're a Manus user

Q: How does the integration actually work?
A: Manus uses ElevenLabs' APIs to generate speech, clone voices, and transcribe audio on your behalf. The workflow is managed by Manus; ElevenLabs processes the audio. Generated outputs are returned within your Manus session. Manus only accesses 3rd party connectors you have already authorized. No new permissions are granted automatically.

Q: Do I need an existing ElevenLabs account to start?
A: Yes. You need an active ElevenLabs account to connect it to Manus.

If you're an ElevenLabs user

Q: Does Manus replace my ElevenLabs dashboard?
A: No. Manus builds a conversational interface that connects to your ElevenLabs account on the backend. Your existing history and voices stay untouched; ElevenLabs keeps managing generation and billing as usual.

Q: Does Manus access my audio data?
A: Manus reads and writes audio through ElevenLabs' API to power features like transcription and generation. Audio data is processed according to ElevenLabs' data retention policies. Manus only accesses 3rd party connectors you have already authorized. No new permissions are granted automatically. Review ElevenLabs' privacy policy for details on how your data is handled.

Bring Sound into Every Manus Task

Most work tools are silent. The ElevenLabs connector changes that, letting anything you create in Manus speak. This is part of a broader direction for Manus: bring the tools and services behind your work into the same place where the work actually happens.

Ready to build something amazing?

Try Manus today →

Supabase Stores the Data Behind Your App

Every digital product needs a database to store its working information: customer accounts, orders, subscriptions, product activity, and error logs. Supabase is a popular database platform where many apps keep that information.

If your product was built using Supabase, all of its live data already lives there. However, accessing that information or making structural changes usually requires technical tools or help from an engineer.

Manus Can Work Directly with Your Supabase Project

The Supabase connector gives Manus permissioned access to your Supabase project.

Once connected, you can ask a question in plain English. Manus will find the right project, retrieve the relevant data, analyze it, and explain the answer.

Work with Existing Data or Build with Supabase MCP

Use Manus to query and operate data already in Supabase, or use Supabase MCP as the backend while Manus builds a new app.

Move Legacy Data into Supabase in One Session

You need to move customer records from an old CRM spreadsheet into your Supabase database. Normally, an engineer would need to clean the data, write a script to map the columns, and handle the import.

With the Supabase connector, you ask Manus:

“Migrate this attached CRM export into our Supabase customer table. Clean the data, remove duplicates, and show me the proposed schema changes before you begin.”

Manus inspects the file and the current Supabase setup. It proposes the necessary schema updates, waits for your approval, and then executes the migration. It handles the data mapping and validates the result.

You then ask:

“Check the system logs to ensure the new records are syncing correctly with our email marketing tool.”

Manus reviews recent system activity and confirms the integration is working. The team moves from a raw data file to a live, verified database update in one session.

Sync, Enrich, and Build with Your Supabase Data

When you connect Supabase to Manus, you are giving an AI agent the ability to act as your data engineer. Here is how teams are using the connector to move faster:

Sync Data from Other Platforms

Instead of manually exporting spreadsheets or paying for expensive integration tools, you can ask Manus to synchronize data from external sources directly into Supabase. Teams use this to pull in Slack messages for volume tracking, sync ad performance from Meta, or collect daily judicial publications and filter out irrelevant records.

Enrich Your Existing Records

If your database is missing details, Manus can find them and fill in the gaps. Users routinely ask Manus to scrape web sources for missing product specifications, classify sales leads based on their digital footprint, or use AI to generate image descriptions and canonical categories for e-commerce catalogs.

Construct AI Knowledge Bases

Building a system that can answer questions based on your own documents requires a vector database. You can instruct Manus to acquire documents, process them, generate the necessary embeddings, and store them in Supabase so your AI features have the context they need.

Repair and Audit Complex Systems

When things break, finding the cause often requires digging through the database. Education platforms use Manus to identify missing practice-test questions and correct the data. Other teams ask Manus to audit payroll systems or e-commerce repositories, find the security or integrity issues, and apply the fixes.

Manus Handles Database Work from Query to Deployment

• Execute SQL and migrate schemas: Manus can write and execute SQL queries, apply schema migrations, and design database structures based on your instructions.
• Deploy Edge Functions: Build and deploy serverless compute functions directly to Supabase without leaving the conversation.
• Check security and performance: Surface Supabase recommendations that may help protect data or improve how the system runs.
• Investigate errors: Review recent system activity to understand failed requests, sign-in problems, data errors, or other unexpected behavior.

Try these prompts:

Understand your product data:

“Show me my Supabase projects and find the one with customer subscriptions. Compare cancellations from the last 30 days by plan tier.”

Improve your setup:

“Design a new schema for tracking user feedback, including the necessary SQL migrations, and show them to me before applying.”

Connect Supabase and Describe the Outcome

1. Open Settings → Connectors in Manus.
2. Select Supabase and authorize the organization you want to use.
3. Start a Manus session with the Supabase connector enabled.
4. Describe the result you need.

Available Now in Manus Connectors

The Supabase connector is available through Manus Connectors. You need a Supabase account and permission to access the projects you want Manus to use. The actions available to Manus depend on the access you grant and the features included in your Supabase plan.[2]

Common Questions

Q: Do I need technical database knowledge?

A: No. You can describe the question or outcome in plain English. Manus can find the relevant data, explain the result, and write the necessary SQL.

Q: What can Manus access?

A: Manus can work only with the Supabase organizations and projects covered by the access you authorize. You can revoke the connection from your Manus integrations settings.[1]

Q: Do I need to build my app on Manus to use this?

A: No. The connector works with any Supabase project. You can use it to analyze data, migrate records, or operate the backend for apps built anywhere.

Bring Live Product Data into Every Manus Task

The Supabase connector gives Manus access to the live product context behind your questions. This makes it easier to understand what is happening, design structural improvements, and verify the result without coordinating the work across several tools.

This is part of a broader direction for Manus: bring the data and services behind your work into the same place where decisions and actions happen.

Most AI Mistakes Happen Because No One Reviewed the Approach

You write a prompt. You wait while Manus builds a full website. When it finishes, you realize it picked the wrong layout, missed the payment integration, or targeted the wrong audience. You spend the next hour correcting it message by message.

The problem isn't capability. It's alignment. Plan Mode introduces a review step before execution begins. Manus pauses, runs a feasibility check, and surfaces the approach as a structured document. You read it, edit it, and approve it. The mistake gets caught before any code is written, before any time is invested in the wrong direction.

Manus Thinks First, Then Builds

The idea behind Plan Mode isn't new. It's the same principle behind project kickoff docs and design reviews: align on the work before the work begins.

Plan Mode is manual. It stays out of your way until you call it. On web, type / and select Plan. On mobile, tap + and choose Plan Mode. Once active, Manus evaluates your request. If context is missing, it asks clarifying questions. If the path is clear, it generates the plan immediately: a full Markdown document with goals, steps, and constraints.

You can click into the document and rewrite anything. You can ask Manus to edit it for you. When you hit Confirm, that plan becomes the source of truth. Manus follows it precisely.

One critical detail:

Manus will not start building until you confirm or dismiss the plan.

This prevents any unintended changes to your website, slides, or video. You stay in control.

Enable It Anytime, Even Mid-Task

Plan Mode isn't limited to the start of a task. You can enable or disable it flexibly at any point during the process. If Manus is already building and you want to pause, regroup, and plan the next iteration, just activate Plan Mode mid-task. Manus stops, generates a plan for the next phase, and waits for your approval before continuing.

This makes it natural to work in stages. Let Manus finish phase one, then turn on Plan Mode to scope phase two before it begins.

Plan Mode Is Built for Development

Plan Mode is most powerful in development scenarios. You describe what you want to build, and Manus uses the planning process to surface what you haven't considered yet: edge cases hidden in your requirements, business logic conflicts buried in the flow, and assumptions that haven't been clarified.

In practice, the planning process simulates the development process itself. Manus walks through your requirements the way a developer would during a technical review. The difference is that catching these issues in a plan costs far less than catching them after the build is live.

This is where the real value sits: avoiding misalignment that leads to over-investment in the wrong direction, and surfacing hidden edge cases early so you spend less time polishing during acceptance. You see the plan in minutes, not the build in hours. Nothing changes until you confirm. And when you want Manus to revise what it already shipped, just edit the plan again.

One Prompt Is All You Need

With Plan Mode, a single sentence is enough to kick off a full website build. You don't need to front-load every detail into your prompt because the plan fills in the gaps for you. Manus takes your brief input, expands it into a structured approach, and presents it for review. If you want to add more context, you can. But you don't have to. The planning process removes the pressure to prompt-engineer perfectly.

Here's what that looks like in practice:

"Build a landing page for my SaaS product."

That's the entire prompt. Plan Mode expands it into target audience, hero copy angle, component stack, and integrations (Stripe, analytics). You glance at the plan, edit one line to prioritize developer signups over enterprise demos, and confirm. Manus builds the right site the first time.

"Build a task manager with user accounts and a database."

One sentence. Plan Mode maps the data schema, auth flow, and API routes. You spot a flat table where you need relational joins. You fix the schema in plain text. The app launches with the right architecture.

"Add a testimonial section and change the CTA on my live site."

Plan Mode shows exactly where the new section goes and which components it touches. You confirm. The update ships without breaking the existing layout.

"Deploy my app and set up analytics tracking."

Plan Mode lists the hosting config, the events it will instrument, and the pages it will tag. You add one missing conversion event. The deploy goes out with complete tracking from day one.

Beyond Plan Mode: More Ways to Reduce Trial and Error

Plan Mode is one part of a broader set of tools designed to lower the cost of iteration. If the latest changes to your website don't work out, you can roll back to a previous version. If you want to experiment without risk, make a copy of your site and try things on the duplicate. And during the planning stage itself, you can use branches to explore multiple different requirement paths before committing to one. Together, these give you room to iterate without fear of breaking what already works.

How to activate plan mode

1. Type / and select Plan from the command menu.
2. If Manus needs more context, answer the clarifying questions.
3. Review the generated plan.
4. Edit directly or ask Manus to revise it.
5. Click Confirm. Manus builds exactly what you approved.

Available Now on Web and Mobile for All Users

Plan Mode is live today for every Manus user on both web and mobile

Common Questions

Q: Does Plan Mode activate automatically?
A: No. It's manual-only. Your quick tasks stay quick.

Q: How do I turn it on?
A: Type / and select Plan.

Q: Can I enable Plan Mode in the middle of a task?
A: Yes. You can turn it on at any point. Manus will pause, generate a plan for the next phase, and wait for your confirmation before continuing.

Q: Can I edit the plan multiple times before confirming?
A: Yes. Rewrite it as many times as you need until you're satisfied with the approach.

Q: Will Manus start building before I approve?
A: Never. Manus will not execute until you confirm or dismiss the plan. Your existing website, slides, and video remain untouched until you say go.

Q: Can I change the plan after I've already confirmed it?
A: Yes. Edit the plan again later in the session and Manus will adapt its work.

Q: Why not just let Manus start building immediately?
A: Complex builds involve hidden edge cases and unstated assumptions. Reviewing a plan takes minutes. Catching a misalignment in text is vastly cheaper than catching it hours later in a deployed site.

Generate, edit, and download .pptx files — without conversion.

Manus now generates PowerPoint files directly. Not converted from another format. Not exported from a web preview. The .pptx is the output from the start — with editable charts, structured tables, and AI-generated visuals that open perfectly in Microsoft PowerPoint or Google Slides.

Build PowerPoint Decks Directly in Manus

Manus has always been able to create presentations. Until now, those decks were generated in a web-based format and converted to .pptx on export. Now, Manus authors a native .pptx from the first slide to the last.

This means every element in your deck is structurally real: charts backed by data tables you can click into and edit, tables with proper cells that support merging and resizing, and layouts that follow standard slide masters. It also means you can now directly download any deck built on Manus as a fully functional PowerPoint file, rather than hoping an export step preserves what you see on screen.

Turn Anything Into a PowerPoint Deck

PowerPoint mode works with every input Manus already supports — but now the output is a native .pptx you can download directly.

• Spreadsheet data → deck.

  Connect Google Sheets through the Workspace connector, point Manus at your data, and ask for a presentation. Manus analyzes the numbers, identifies trends, builds editable charts, and outputs a .pptx where every visualization retains its underlying data. Update a cell in PowerPoint and the chart adjusts.

• Research topic → deck.

  Describe a subject and Manus conducts

  wide research

  across multiple sources, synthesizes the findings, and generates a slide deck with properly cited content and data-backed visuals → all in native PowerPoint format.

• PDF or document → deck.

  Upload a report, whitepaper, or brief. Manus extracts the key points, structures them into a narrative, and builds a presentation you can hand directly to your audience.

• URL → deck.

  Point Manus at a webpage → a competitor's site, a product page, an article → and get a structured analysis or summary packaged as a .pptx.

• Meeting notes → deck.

  Paste raw notes or a call transcript and Manus turns them into a polished recap deck ready for follow-up distribution.

Edit Your Data Without Leaving Manus

Because PowerPoint mode generates structurally real charts, you can edit them right inside Manus before you ever download the file. Click any chart and a toolbar appears with two controls: Elements and Edit Data.

Elements lets you toggle chart components on and off → axes, data labels, gridlines, legend → so you can customize how your chart looks without rebuilding it. Turn on data labels and every bar segment instantly displays its value. Turn off gridlines for a cleaner look. Each toggle updates the chart in real time.

Edit Data opens a spreadsheet overlay directly on top of your slide. The left side shows your data table → rows, columns, the actual numbers behind the chart. The right side shows a live preview. Change a value in any cell and the chart redraws instantly: the bar grows, the label updates, the proportions shift. When you are done, hit Confirm and the chart on your slide reflects the new data. No round-trip to Excel. No re-importing a CSV. Just click, edit, and see the result.

This is what native PPTX makes possible. The chart is not a flat image → it is a real object with real data behind it. Edit it in Manus, or download and edit it later in PowerPoint. Either way, the data stays live.

How to Use PowerPoint Mode

Open Manus and select the Slides tool. In the mode dropdown, choose PowerPoint (you will see it alongside Standard and Image). Describe your presentation topic, upload data, or connect Google Sheets for live data access. Manus outlines the deck, designs the slides, and generates the .pptx. You can edit charts and tables directly in the Manus viewer before downloading, or download immediately and continue in PowerPoint.

Availability

PowerPoint mode is available today in Beta for users on the 1.6 and Max models.

Common Questions

Q: What is the difference between Standard, Image, and PowerPoint mode?
A: Standard is fast and versatile for general decks. Image creates high-impact visual slides powered by GPT Image 2. PowerPoint generates a native .pptx file optimized for data-dense business presentations with editable charts and tables.

Q: Can I pull data from Google Sheets into my presentation?
A: Yes. Connect your Google Workspace account, point Manus at a spreadsheet, and it will analyze the data and generate a PowerPoint deck with editable charts based on your numbers.

Q: Will my charts be editable after I download the file?
A: Yes. Every chart retains its embedded data table. Click any chart in Microsoft PowerPoint or Google Slides and edit the numbers directly.

Q: Can Manus research a topic and create the deck in one step?
A: Yes. Describe your topic and Manus will conduct research, synthesize findings, design slides, and output a native .pptx → all in a single task.

Q: How is this different from how Manus handled PowerPoint before?
A: Previously, Manus generated slides in a web format and converted them to .pptx on export, which could cause formatting issues. PowerPoint mode skips the web layer entirely and authors the .pptx directly, preserving chart editability, table structure, and layout fidelity.

The best presentations combine sharp thinking with clean delivery. PowerPoint mode means you no longer have to choose between the speed of AI generation and the reliability of a native file. Research, analyze, design, and deliver → all from one place

Stop publishing manually — your site updates the moment it is ready

Building a website with Manus is fast. You describe what you want, Manus writes the code, and a working version appears in seconds. But getting that version onto your live URL has always required a manual step: open the publish menu, confirm your settings, click Publish. Across an afternoon of iteration — fifteen small tweaks, fifteen manual publishes — that step accumulates into the thing that actually slows you down.

With auto-Publish, a single toggle inside the publish popover tells Manus: every time a build succeeds, deploy it. No confirmation needed for your live site stays in sync with the latest version you accepted, automatically.

Every successful build goes live instantly

When you enable the toggle, Manus treats every successful build as a deployment event. The moment your site finishes building, it goes live at your public URL. If you are iterating quickly — adjusting copy, swapping images, tuning layouts — each change lands on your domain as soon as it is ready. You stay in the conversation. The site stays current.

The toggle is off by default. This is intentional. Not every iteration is meant for the public. You might be experimenting with a layout you are not sure about, or testing a color scheme before committing. Auto-Publish is designed for the moments when you know the direction and want to move fast, when the distance between "looks good" and "it's live" should be zero.

You can turn it off at any time from the same publish popover. Nothing changes about your existing workflow unless you opt in.

You know the direction — now you just need it shipped

Picture this. You are building a portfolio site for a freelance photography business. The first session gets the structure right: a grid gallery, an about page, a contact form. You publish it manually, share the link with the photographer, and get feedback overnight.

The next morning, you sit down with a list: swap the header font, add a client testimonials section, compress the gallery images, update the footer with social links, and fix the mobile nav breakpoint. You know exactly what you want. You are not exploring — you are finishing.

This is where you toggle Auto-Publish on. Each change Manus completes goes live immediately. The photographer can refresh the page between your messages and watch the site evolve in real time. By the time you finish your coffee, the site is done and live, without a single trip to the publish menu.

Auto-Publish is for sessions where you have already decided to ship. It removes the ceremony so you can focus on the work.

One toggle, inside the publish popover

1. Open your WebDev project and click the Publish button in the top-right corner.
2. In the publish popover, scroll to the bottom.
3. Toggle Auto publish when it is ready to on.
4. Continue building. Every successful build will now deploy to your public URL automatically.

To stop auto-publishing, open the same popover and toggle it off. Your next build will wait for manual confirmation before going live.

Go further: queue your changes and walk away

If you already know the full list of changes you want, you can queue multiple messages while Manus is still working on the current one. Combined with Auto-Publish, this means you can drop your entire to-do list into the chat, close your laptop, and come back to a finished, live site. Manus processes each request in order, builds the new version, and deploys it — all without you being present.

Availability

Auto-Publish is available to all users on web, iOS, and Android starting today.

Common Questions

Q: Is Auto-Publish on by default?
A: No. The toggle is off by default. You opt in when you are ready to ship continuously.

Q: What if I share my URL with a client — will they see half-finished work?
A: Only completed, successful builds are published. If a build fails or is still in progress, your live site remains on the last stable version. Your client always sees finished work.

Q: Can I cancel a publish that is already in progress?
A: Yes. The interface shows a cancel option during deployment if you need to stop it.

Q: Does this work on mobile?
A: Yes. The Auto-Publish toggle is available in the publish popover on both iOS and Android.

We built Manus to handle complete work — not just write code, but deliver finished products. Auto-Publish removes the last manual step between creation and delivery: the moment you accept a change, it reaches your audience.

Work is not linear — your AI shouldn’t be either

A single research report can become API documentation, a user manual, a flowchart, and a slide deck. An initial idea radiates outward into many parallel tasks of different types. Yet most AI tools force all of that into one forward-moving thread, where every new direction dilutes the context the model has already built. The real cost of starting over is not time — it is losing the accumulated understanding.

Branch treats that understanding as a reusable asset. Click the branch icon beneath any message and Manus opens a parallel session that inherits everything up to that point — your instructions, your files, and the full history. The original stays exactly as it was. The new branch becomes a separate space where you can take the same starting context somewhere new, without contaminating what came before.

The result: you can spend the same context many times without depleting it. Each branch keeps its own clean focus, so the work you do in one direction never confuses another.

Your context stays intact — you just iterate from it

When you branch, the new session inherits your files, your instructions, and the full conversation history up to that point. You don’t need to re-upload or re-explain anything. Your original session stays untouched, so the work you’ve already done is not at risk. And because each branch carries its own isolated copy of the context, the directions you explore don’t bleed into each other — Manus stays focused on what that particular branch needs, which keeps output more accurate and focused. A "Branched from" breadcrumb ties every new session back to its origin, so you can always trace where a branch started and navigate between them as your workspace grows.

How you might use Branch

Picture these situations.

You just finished an online meeting and had Manus put together the notes. The summary looks good, but now you need several things from it: a to-do list for the team, a progress update for your manager, and a slide outline for next week’s review. If you try to produce all of those in the same session, the conversation grows long and Manus’s focus can drift. Instead, you branch from the point where the meeting notes were finalized. Each branch focuses on one output, while the notes themselves — the shared starting point — stay intact in every direction.

Or say you’re running a market research project. The research itself is an ongoing effort: collecting sources, organizing data, building up a picture over time. But midway through, you want to produce an investor brief based on what you have so far, and separately, an internal analysis report. You don’t have to wait until the research is complete. Branch from the current point, and each branch inherits everything gathered so far. You produce phased outputs while the main thread keeps moving forward.

Or imagine you’ve spent a session writing and refining a blog post. By the time it’s done, Manus has a deep understanding of the narrative, tone, and key points. Now you want to turn that into a video script. If you keep going in the same session, blog edits and script iterations will get tangled. Branch out, and you can focus the new session entirely on pacing, voiceover, and shot direction — while the script stays grounded in the original piece because it inherited the full context.

There’s also the running-record pattern: one ongoing session that acts like a living project log, accumulating notes and context over time. Whenever you need a specific deliverable, you branch from the current point to produce it — a plan, a summary, a presentation — and the main thread continues recording and growing. Your “main line” stays a complete project memory, while each branch turns one moment in that record into a finished output.

The common thread: your work doesn’t have to follow a single straight line. Branch lets you say “I want to go in a new direction from here” at any point, without starting over.

How to create a branch

1. Hover over any previous message in your conversation.
2. Click the branch icon that appears beneath it.
3. Wait a moment while Manus spawns the parallel session — you'll see a "Session branched" confirmation.
4. Continue in the new branch, which already carries all of your prior context. Repeat from the same message to explore as many parallel directions as you need.

Availability

Session Branching is available to all users in standard Manus chat sessions. It is not currently supported in Web Builder sessions — branching applies to regular conversations and tasks, and support for additional session types may follow.

Common questions (FAQ)

Q: Does branching delete or change my original conversation?

A: No. Branching creates a separate, parallel session. Your original conversation stays intact and unchanged.

Q: Does the new branch keep my files and context?

A: Yes. Everything up to the branch point — instructions, uploaded files, and history — is inherited by the new session automatically.

Q: Can I create more than one branch from the same message?

A: Yes. You can branch from the same point as many times as you like to explore multiple parallel directions from one shared context.

Q: Can I branch from a branch?

A: Yes. Any standard session can be branched, so you can explore deep decision trees without losing your place.

Q: Can I branch a Web Builder session?

A: Not yet. Branching is currently supported in standard Manus chat sessions and is not available for Web Builder. We may extend it to more session types in the future.

What comes next

Branch is a step toward AI that mirrors how you actually think. You don’t hold a body of knowledge and ask one question — you ask many, in many directions, simultaneously. By turning every conversation into a reusable foundation you can build on without depleting, we want the question "what could this become?" to cost nothing but curiosity.

Mobbin + Manus: Turn Proven App Designs Into Finished Work

Building an app, a deck, or a design critique from a prompt is no longer the hard part. The harder question comes right before that: what should the checkout screen look like, where does the pricing toggle go, how do the apps people already trust handle their onboarding?

Mobbin, a UI design reference library with over 600,000 real screens from shipped products, gives you proven foundations so you start from what works rather than from a blank slate. Once connected, Manus reads the screens, flows, and sections Mobbin returns and turns them into the finished thing, whether that is a coded app, a stakeholder deck, or a marked-up critique.

See What Works. Manus Builds It.

On its own, a research library shows you the screens, flows, and sections that real apps have shipped. What it does not do is build the thing. You still have to take what you learned and turn it into a coded app, a deck, an analysis, or a critique somewhere else. That handoff is where the connector pays off. Manus picks up exactly where the research ends.

4 Ways to Put the Mobbin Connector to Work

The connector exposes three search tools inside Manus: searching single screens, analyzing multi-step user flows, and reviewing specific website sections. The point is not the tools themselves, but what happens when the references they return flow straight into the work. Here is how that plays out across a slide deck, a mobile app, an industry analysis, and a design you already have.

1. Turn KYC Research Into a Stakeholder Deck

If you are presenting to a product or compliance team, the hard part is not the slides, it is the evidence behind them. Saying a sign-up flow should feel reassuring lands better when you can show how the apps people already trust actually handle it.

• Researches real KYC screens from leading global fintech and cryptocurrency platforms through the connector, reading what each one does.
• Builds the deck with its Slides Generation capability, annotating each screenshot with the pattern, its strength, and its trade-off, and closing on recommendations tied back to the apps that use them.
• Delivers a deck that exports to PowerPoint, PDF, or Google Slides, so it is ready to present, not just to read.

Prompt: "Create a stakeholder-ready slide deck titled 'How the top apps in fintech handle KYC onboarding,' using the Mobbin connector for research. Research how around ten leading fintech apps handle KYC onboarding, capture representative screenshots, and identify the key patterns, strengths, and trade-offs. Build a deck with a title slide, individual slides for each app or pattern with annotated screenshots, a synthesis slide, and a recommendations slide. Include speaker-note rationale and Mobbin deep links."

2. Research Patterns and Build the App in One Pass

The same handoff works for mobile. Instead of generating an app from a blank slate, you can have Manus study how similar apps handle their core screens first, then build on top of those findings.

• Researches camera app patterns through the connector, looking at how top-tier photography and social media applications handle the viewfinder, the capture controls, and the gallery grid.
• Builds the app on top of that research, generating the artwork and a working camera screen with a film-stock filter dial, a film rolls tab with a develop timer, unlockable skins, and an achievements tab.
• Delivers a working prototype, not a moodboard, that you publish straight from Manus and open on your phone in Expo Go, the testing app that runs a build before it goes to any store.

Prompt: "Using the Mobbin connector, research the best UI and UX of photo camera apps. Then build a cross-platform mobile app: a retro keychain digicam that captures lo-fi photos with grain, auto-flash, an orange date stamp, twelve vintage filters, and a small washed-out LCD viewfinder. Add a blind-box onboarding that unlocks seven collectible camera skins, plus a gallery, sharing, achievements, and a film roll develop-delay mode. Make it charmingly imperfect and uniquely differentiated."

3. Turn a Category Into a Pattern Atlas

Not every job ends in a build. Sometimes the deliverable is a decision, and you need to see how a whole category handles one journey, across many apps, with the numbers behind what is common and what is rare.

• Researches the top ten apps through the connector, then runs Wide Research to study every flow across all of them in parallel, dozens of analyses at once.
• Builds the findings into two outputs: an Excel spreadsheet with one row per app and flow via its Data and Spreadsheets capability, and a hosted site via its Web Development capability that groups the apps by pattern.
• Delivers a category teardown with receipts, including how often each pattern appears, without downloading a single app.

Prompt: "Build a 10-competitor pattern atlas for the language learning category using the Mobbin connector. Use the connector to pick the top ten apps, then use Wide Research to analyze their onboarding, checkout, and paywall flows. For each app and flow, record the pattern type, the number of steps, two to three notable UX choices, a representative screenshot, and the Mobbin deep link. Compute the frequency of each pattern per flow. Deliver a spreadsheet with one row per app and flow, and a visual website that groups the apps by pattern, shows the frequency stats and screenshots, and ends with a 'So what' section of the top takeaways tied to the evidence."

4. Critique a Figma Mockup, Then Mark Up the Fixes

Sometimes you already have a design, but you need to know if it holds up against what is shipping now. Instead of relying on subjective feedback, you can have Manus compare your draft against real, proven pricing pages.

• Researches pricing sections through the connector, pulling real references from industry-leading design, project management, and web development tools.
• Builds a structured critique against your Figma mockup, calling out what to keep, like a clear monthly and annual toggle, and what to fix, like a missing comparison table, no FAQ, and no social proof near the toggle.
• Delivers notes you can act on in plain language, and a short follow-up turns those written notes into a visual:

Prompt: "Using the Mobbin connector, research the best practices of pricing page design, then critique this Figma mockup: [your Figma link]. Search the connector for web pricing page sections with plan comparison, review my mockup against what you find, and deliver a report covering what works and the specific areas to improve."

Follow-up prompt: "Using image generation, annotate these recommendations onto the Figma screenshot in a hand-drawn style, and crop out the footer."Manus marks the fixes directly onto your design, circling the plan to highlight, pointing an arrow to where a 'Most Popular' badge should sit, and noting where to add the comparison table, the FAQ, and the customer logos. You end up with both a written critique and a marked-up image your team can act on at a glance.

Setting Up Your Mobbin Connection

Connecting Mobbin to Manus requires a paid Mobbin plan (Pro or Team) and takes just a few clicks.

1. Open your Manus workspace and navigate to the Connectors tab.
2. Select the Mobbin connector and click "Connect."
3. Follow the prompts to sign in and authorize secure access to your Mobbin account. Your credentials and data remain under your control. Manus only accesses the Mobbin library when you explicitly authorize it for a task. The connector is strictly read-only. It cannot edit, delete, or manage your saved collections. You can monitor or revoke access at any time from your Manus Connectors settings.

Common Questions

Q: What do I need to connect Mobbin to Manus?
A: A paid Mobbin plan (Pro or Team). Once that is in place, you connect and authorize access from the Connectors tab in your Manus workspace.

Q: Can Manus change or delete anything in my Mobbin account?
A: No. The connection is strictly read-only. Manus can read screens, flows, and sections, but it cannot edit, delete, or manage your saved collections, and you can revoke access at any time.

Q: What can Manus actually build from the research?
A: Whatever you ask for, including a coded app, a stakeholder slide deck, an industry analysis with a spreadsheet and a hosted site, or a marked-up critique of a design you already have.

Q: Which apps does Manus pull references from?
A: It surfaces real, shipped screens from leading apps across the categories you research. References returned through the connector link back to their source for verification.

Q: Who can use this connector?
A: Any Manus user who connects an eligible paid Mobbin plan. Manus only accesses the connector after you explicitly authorize it for a task.

Stop Guessing What Works

On its own, Mobbin answers the design question, and Manus on its own can build whatever you describe. Connecting them puts the research and the work in one place, so the proven screen you found becomes the app you ship, the deck you present, or the critique you hand back, without a handoff in between. You decide what to make and which patterns to trust, Mobbin supplies the credible references, and Manus turns them into the finished thing. Your data stays yours, the connector reads from the Mobbin library only when you authorize a task and never writes to it, and you can revoke access at any time. Connect your account and start building from what already works.

Manus only accesses 3rd party connectors you have already authorized. No new permissions are granted automatically. AI-generated outputs may contain errors. Users should independently verify any data, screenshots, or recommendations before finalizing designs or deployments.