Teamily AI - Your personal agent, agent teams & people in one messenger
by•
The human-agent platform where people and always-on agents get work done together in one unified, context-aware space. Unlike single-user personal agents, Teamily AI enables multi-user, multi-agent teamwork. Built on an AI-native messenger, it connects people, agents, docs, meetings, calls, workflows, and business systems to automate office work and customer service. Agents become real teammates with persistent memory, proactive intelligence, and continuous self-improvement. Free to get started.
Replies
👋 Hi Product Hunt! I'm Salman Avestimehr, Co-founder of Teamily AI and Professor at USC. Excited to introduce Teamily AI to this community! 🚀
Having spent years researching collaborative and distributed AI, I've become increasingly convinced that the next major breakthrough in AI won't simply come from making individual models smarter. It will come from enabling humans and AI agents to think, collaborate, and continuously learn together.
Today, AI is largely built around one person interacting with one agent at a time. But real world intelligence, creativity, and productivity emerge through teamwork, where people share knowledge, build on each other's ideas, and solve problems together.
That's the fundamental shift we're pursuing at Teamily AI: moving from individual intelligence to collective human-agent intelligence.
We're building the foundations for this future through shared context, persistent memory, multi-agent collaboration, and continuous learning, all brought together in an AI-native messenger where humans and agents work as one team.
I believe this will fundamentally change not only how we work, but also how we create, discover, and innovate.
Would love to hear from the Product Hunt community: What becomes possible when AI is no longer just your personal assistant, but a true teammate?
@salman_avestimehr1 nice！
@salman_avestimehr1 @auula_ Leon you're everywhere today 😂🙌 thank you for cheering on the whole team, it genuinely means a lot on launch day!
@salman_avestimehr1 I also share my vision here: https://x.com/AidenChaoyangHe/status/2117613537929703708
@aidenchaoyanghe nice!
@salman_avestimehr1 Do you think AI agents will work better when they share knowledge instead of working alone?
@shivam_kushwaha16 Definitely! Just like us, agents can achieve a lot more when they share knowledge, build on each other's work, and learn together. And the really exciting part is how that collective intelligence compounds over time.
@salman_avestimehr1 @shivam_kushwaha16 Shivam +1 to Salman on this! From the design side I'd add one thing: sharing only works if people can see what got shared. If an agent learns something in one conversation and uses it in another, the people in the room should be able to tell where it came from. Otherwise shared knowledge starts to feel a bit like gossip 😅 That visibility layer is a big part of what I'm working on.
@salman_avestimehr1 Could it pull themes from a batch of user interview notes???
@salman_avestimehr1 @amrita_prajapati2 Such a designer question Amrita 😄 Yes, drop the interview notes into a group, tag an agent and ask what keeps coming up across them. The best part is what happens next: the team can push back on its themes right there in the thread ("wait, that's one loud user, not a pattern"). Synthesis gets so much better when it gets argued with, not just read 🔥
@salman_avestimehr1 What happens to agent tasks when the person who started them leaves the group???
@vipul_kumar1280 if they don't want the tasks to continue (without them being still in the group), they should take their agents off with them and abort the tasks.
@salman_avestimehr1 That makes sense! So the agents and their tasks are tied to the person who started them. Thanks for clarifying!
@salman_avestimehr1 @vipul_kumar1280 Vipul good one, and Salman's answer is how it works today. The design question I keep chewing on is the in-between: if you leave but the task matters to everyone else, should it quietly stop, or get handed to someone still in the room? Our agents are built to keep going on long tasks, so "who owns this now" becomes a real moment to design, not an edge case 👀
👋 Hi Product Hunt! I'm Xavier, Product Lead at Teamily AI.
Teamily AI is the Human-Agent Platform: one shared space where your team and your AI agents work as one team brain.
Think: your group chat, but your AI agents are teammates with memory.
Huge thanks to our hunter @rohanrecommends for hunting us 🙏
Create a group, invite your teammates, add as many agents as you need, and @ mention any of them. They see what you share, create real work, and keep going after you log off.
Before Teamily AI, I was a product manager at Tencent, then founded an AI agent SaaS startup. Building agents for teams, I kept seeing the same problem: each agent was smart on its own, but none of them shared context with each other or with the people around them. Our founders bring the other half: @aidenchaoyanghe holds a USC PhD in distributed machine learning and previously worked at Tencent, and @salman_avestimehr1 is a USC professor and IEEE Fellow. We've spent years on how people connect, and how machines learn together.
Here's the problem we kept running into. Every new AI tool made each of us faster, and our teams more scattered. Someone drops a link in the group chat, three people ask "wait, what's the context?", and one person becomes the copy-paste layer between five AI tools.
We have smarter individuals, but scattered teams.
So we built Teamily AI to put people and agents in the same room:
🎨 Understand anything. Photos, videos, links, and docs dropped into the chat.
🛠️ Create real work. Websites, slide decks, videos, and dashboards, right in the thread.
🤝 Work as a team. Multiple agents split a task, hand off, and push back on each other, while you steer.
🧠 Share one living memory. It carries across groups, sessions, and agents. A new teammate @ mentions an agent and gets the full story of last month's pricing decision.
🔌 Act in your apps, always on. Connect Gmail, WhatsApp, GitHub, and X. Agents send, post, and follow up while you sleep.
📈 Get better with your team. Each agent has its own identity and persona, and learns your tone, terms, and standards over time.
We call this Group Intelligence: humans and agents getting smarter together. Our goal is to give every person and every team their own AI teammates that grow with them.
It's still early, and there are rough edges. That's exactly why we're here.
I'll be in the comments all day. Which group chat would you drop an AI agent into first?
@0x47xavier yeah exactly, group chat is already where all the context lives. really interesting idea, congrats on launch!
@0x47xavier @sandy_liusy Exactly Sandy!! The context was already there 😄 We just stopped asking people to carry it somewhere else. Group chat has always been where the real decisions happen, so it felt natural to bring the agents to it, not the other way around 🙌
@rohanrecommends @aidenchaoyanghe @salman_avestimehr1 @0x47xavier Nice！
@aryan_mehta6 Of course, Just ask in the same chat, something like "give me three rough options for X", and you'll get all three right in the thread. You can also @mention a few different agents and have them each take a different angle in parallel, so the team can compare side by side and pick one to develop further.
@0x47xavier @aryan_mehta6 Aryan yes, just ask in the same chat! Like Xavier said, you can also tag a few agents and have each take a different angle. The fun part is the comparing, three rough options side by side in one thread are so much easier to react to than three separate chats 😄
Congrats for the launch! Teamily is most interesting when the agents act on the discussion instead of only summarizing it.
@brattyaiguy Thanks, RI-Chie, totally agree! Our agents split a task, build things in parallel, and hand results back to the group while you steer. What would you want an agent to take on first?
@brattyaiguy Totally with you RI-Chie!! A summary is nice, but the moment an agent actually picks up the next step is when it starts to feel like a teammate. Ours is honestly a bit of a workhorse, it keeps going on a long task even if nobody replies for a while. The design work now is making sure it acts on the right part of the discussion, not just the loudest one 🙌
Wow!! congratulations on the launch. This is a great build. Quick question how does usage limits and credit work with collaborations?
@richatsealedvault Thanks for the kind words! Are you thinking of a teammate tagging an agent in a shared chat? I’ll check whose credits that uses and whether the usage limit is per person or shared by the group.
@michellegv Thanks, I was just wondering how that works not necessarily teammate tagging an agent. Exactly as you will investigate: whether the usage limit is per person or shared by the group.
@michellegv my questions are very well answered. thanks.
@richatsealedvault Thanks so much, Richard! Great question. In a shared group, you choose how usage is covered:
1. Everyone uses their own plan, so each member's usage counts against their own credits.
2. The group owner covers it, either treating the whole group or sharing a pool with everyone.
If you are the group owner, you can switch between these anytime. Teamily AI is free to start, and you can earn extra credits through daily activity, usage leaderboards, and other events. PH folks also get 50% off the first month of Plus with code PHTEAMILY 🎁
@0x47xavier Excellent and thoughtful architecture. That is an impressive build. Thanks for getting back to my question this quickly. Cheers to Teamily AI.
@richatsealedvault Thank you for your question! We have built in very detailed settings here. We have spent several months refining this section, tailoring it to different users so they have as many choices as possible to suit their various lifestyles or working environments. If you feel there are any user use cases we haven't covered, please let us know. We'd love to hear your suggestions! 🥹🙌
When you set up a group chat:
1. We have different options tailored for individuals and teams.
2. The group owner can choose between a members-pay or owner-pays model.
@roxie_ren I am very satisfied the response. This is certainly needed for a team building something together or collaborating on a project. I hope other participants here and potential users will find the answers very informative as well. Thank you Roxie.
👋 Hi Product Hunt! I'm Roxie, Head of Design at Teamily AI.
Really excited to finally share this one. I've always worked where design meets the future, and I've never seen a moment that needs it more. ✨
Every new medium gets designed twice. First by accident, wearing the shape of whatever came before. Then on purpose. Agents are still in round one: mostly one bot talking to one person, a shape borrowed from the last era.
Teamily AI is our shot at round two. People and agents work in the same room, with the same memory. It's also the first step toward what Aiden calls the agentic internet, and that's the part that hooked me.
That asks a lot more of design than nice screens. When should an agent speak up, and when should it stay quiet? What should it remember, and who gets to see it? Those answers shouldn't be buried in a settings page. They belong in every moment of the experience. Honestly, I can't imagine more exciting work for a designer right now.
In my second year here, I worked with our founders to build our identity from scratch, and I've been shaping the story and the experience ever since. We've rebuilt it several times, fast, and we're not done yet. 🚀
Plenty is still open. How proactive is too proactive in a group of ten? Who owns a memory two teams made together? 🙂
Huge thanks to our co-founders, @aidenchaoyanghe and @salman_avestimehr1 , and to the whole team. 🙏
If you try Teamily AI today, tell me what feels like the future and what still feels like just another chatbot. Both are gold. 💚
@aidenchaoyanghe @salman_avestimehr1 @roxie_ren go to it！
@aidenchaoyanghe @salman_avestimehr1 @auula_ haha on it 🙌🙌 thanks for showing up Leon!! your energy in the comments today has been the best part of launch day 😄
Sounds interesting. Are you going to ship mobile apps too?
@aramhakobyan Teamily AI is already available on iOS and Android. You can pick up right where you left off in the group chat.
@aramhakobyan Good news: Teamily AI is already on iOS, Android, web and MacOS desktop. Your groups and agents sync across all of them, so you can pick up the conversation anywhere.
@aramhakobyan We've covered pretty much all mobile apps, and we've also just released an iphone Duo. 🚀We've tailored the UI experience for every suitable mobile app. If you run into any issues or have any feedback while using it on your version of the product, please let me know.
Is there a way to lock a final version before more edits happen?
@sy_jin_3 I’d have to check whether you can lock the draft itself. In the meantime, you can save a separate copy of the approved version and share it in the group chat so everyone knows which one got signed off.
@sy_jin_3 You can ask an agent to save the final version as a shared doc in the group, so everyone has one clear reference while new ideas keep coming in the chat. A proper lock for finished work is something we're looking into, so thanks for the nudge!
@sy_jin_3 In Open Studio, you can see all the previous versions you have edited. Of course, you can lock down a final version, but if you change your mind, you can also find any version you previously locked.
👋 Hi Product Hunt! I'm Aiden, CEO of Teamily AI. I'm excited to introduce the latest version of Teamily to the Product Hunt community and share our vision for its future.
Today, Teamily AI brings people and AI agents together through a multi-user, multi-agent collaboration platform built on a reinvented AI-native messenger. It’s a faster and smarter way to turn conversations into creations and automations, from web apps and slide decks to deep research, dashboards, documents, videos, and more, with connected context, living memory, and self-improving AI agents.
But our ambition goes far beyond a human-agent collaborative platform.
Teamily AI's long-term vision is to build the central platform of the next-generation Agentic Internet, where services and work are delivered by AI agents, no longer wrapped inside traditional web, desktop, or mobile apps distributed through app stores or search engines. Agent-as-a-service (AaaS) will be as accessible as friends on a social network. Just as Google democratized access to information and services in the web era, Teamily AI aims to democratize access to intelligence and services in the Agentic Internet era, making the right intelligence, people, and services equally accessible to everyone.
We’d love to hear your feedback!
@aidenchaoyanghe so proud to be part of this one 💚 the room-of-ten idea is what got me excited from day one. Let's go team! 🚀
Can guests join a group without getting access to all the memory?
@alexia_li Guests cannot access all your memory across your groups. Guests in a group can only see the memory of the group they joined. This is similar to a human-only IM. When a person joins a group, he/she can only see that group's chat history.
But if guests @ one agent in a group, such an agent may disclose some memory it gets from somewhere else.
Overall, we prefer "context not control", but we are also thinking about more fine-grained permission management. We welcome more feedback on interesting use cases.
@alexia_li Alexia adding a little design note to Aiden's answer 🙂 The tricky case is the one he flagged, an agent bringing something in from another room. When it uses a memory, I'd love it to show where that came from, so a guest never has to wonder why it knows that. "Context, not control" works best when the context is visible 👀
shared memory could be the real win here
@emilyrae It’s especially handy when a different teammate picks up the work. They can ask the agent where things left off without waiting for someone to catch them up.
@emilyrae That's exactly how we see it, Emily! Memory belongs to the whole team, not one chat. It carries across groups, sessions, and agents, so context follows the work instead of resetting every time.