The Enterprise Agentic Frontend Stack. Build rich agentic applications that learn from every interaction. Self-hostable and built to scale with your org.
This is the 5th launch from CopilotKit. View more
OpenBot
Launched this week
Build your own enterprise-ready agent workspace that's modular and self-hostable. Designed with a clean architecture and built-in governance. Bring any agent- then choose the models, tools, and infra they run on.
Use as-is, white-label it, or deeply customize it.








Free Options
Launch Team

AWS Activate CreditsCredits that fuel your startup’s next milestone
Promoted







@nathan_tarbert @eli_berman2 what happens when an AI agent encounters a task outside it’s knowledge base ? Can it intelligently escalate to a human ?
@nathan_tarbert @dipanshu_kushwaha5 Yes! OpenBot comes ready with human-in-the-loop approvals. So when your bots come across sensitive tasks, they pause and wait for the user to approve/decline. Also, any user can takeover a bots work whenever they choose, and hand it back when they're ready.
@nathan_tarbert @dipanshu_kushwaha5 @eli_berman2 I really like that OpenBot offers human-in-the-loop approvals for Email Operations! For me, sensitive tasks include sending out email which cannot be undone.
One hundred percent @heidi_spitz1, human approval is built in by default, and it's really important when any destructive action can take place.
Good question @dipanshu_kushwaha5 . A task outside an agent’s knowledge and an action that needs approval are two different cases. Eli covered the approval side; that alone doesn’t mean the agent will reliably recognize every knowledge gap.
You can take over its work and hand it back once you’ve provided the missing context. I would love to dig into this. Is there a particular workflow you’d want to escalate automatically?
@dipanshu_kushwaha5 great question, I can answer this from my own experience - each bot has its own job with instructions, skills, tools and goals (the assigned responsibilities). so when something falls outside that, it doesn’t guess & passes the task to the bot that owns it, and that bot can pass it on again if needed.
if no bot fits or the action is sensitive, it brings you in.. you can approve, decline or take over. workspace works like a real team where everyone knows their role and who to ask
@nathan_tarbert @eli_berman2 @dipanshu_kushwaha5
@dipanshu_kushwaha5 adding the protocol side: pausing for a person is part of the protocol. A run can finish with an interrupt that says what the agent needs, and the next run picks up from the person's answer. So "wait for a human" doesn't depend on keeping a connection open, and any agent that speaks AG-UI can do it the same way.
@markusmxe, appreciate you adding the technical details on how it works!
@nathan_tarbert @dipanshu_kushwaha5 @markusmxe This is coming from the guy who was the mastermind behind AG-UI protocol!
Hey, congrats on the launch!
I'm a huge fan of GrokBot. One of the really active users. I've been using it for three months now. My question here is: What are the limits/usage you have and how quickly do they get hit? Since GrokBot recently adjusted their limits, how many interactions can we do before we hit the highest limits compared to them?
@byalexai No limits on our side. It depends on which AI provider you connect OpenBot to. We're also releasing really soon multi provider support so if you have both an OpenAI and Anthropic subscripcion it's as easy as switching providers once you hit your limits.
@byalexai @guido_vzs Awesome!!
Thanks@byalexai, Aleksandar! Three months of heavy use sounds like a good test case for OpenBot. What workflow would you want to bring over or test first?
I’d be curious to hear how it holds up for the work you’re already doing. We have an OpenBot channel in our Discord community and would love to get your feedback.
https://go.copilotkit.ai/discord-community
Self-hosting plus approvals is the combo that actually matters to me — I've been burned by agents doing too much on their own. Curious how routines work in practice: are they more like scheduled prompts, or can they chain tools and ask for approval mid-run?
@jcbuilds Not scheduled prompts. A routine is a full Bot run on a schedule or a trigger - like a time, or an event or a new email or a pull request. It chains tools the same way it would if you’d asked in chat: reading through apps, using its own browser and shell, handing work to another Bot.
Approvals still apply mid-run. Every action goes through the same gate as a chat turn. Sending, posting, and deleting are governed by your own or your admin’s rules.
When a routine hits one of those, the bot stops and waits for approval. Policies can also tell a routine apart from a person: a rule can refuse a scheduled run something it would allow you live.
A routine also runs as the person who created it, with only their access, so it can’t reach anything they couldn’t. And it’s all on the audit trail, tagged as started by that routine.
More here: https://github.com/CopilotKit/OpenBot/blob/main/docs/routines.md
The open source openbot opens up so many possibilities you can go down on and what you can use in terms of tech without getting locked in.
@alem_tuzlak No vendor lock-in is the key here. You customize it to your entire stack, tools, agents, and infra.
Exactly, @alem_tuzlak, that's one of the biggest benefits. You can bring an existing agent in through AG-UI, connect your tools through MCP, and adapt the workspace to how your team works.
Spent some time this weekend moving tasks from Dots over to OpenBot... The ability to bring my own model and run on my own machine are what sold me on the idea!
@mefogle That's great to hear. How was the experience?
@eli_berman2 Great so far... pro tip for anyone else migrating (should also work with Muse, Duet, etc.) : Ask your current bot to "describe your job so another agent could pick it up" (e.g., triage Github issues) and then feed that prompt to OpenBot. OpenBot actually wrote its own AG-UI connector to drive the task!
@mefogle That's a great tip!! Agents for the win
@eli_berman2 And they don't complain when you tell them to train their successor! 😏
@mefogle, Dots into OpenBot. I would love to hear more.
This is so amazing. I've been using it for a week, and it is the best coworker I've ever had. It does real work in a safe way.
@travis_murdock That's awesome to hear! Where did you find it most valuable?
@travis_murdock, keep the feedback coming, even if it's negative, because that helps us smooth rough edges for everyone else :)
Love the idea of running this yourself. Can you set different approval rules for different actions? I’d be happy to let it organize things, but I’d want to review anything it sends.