I Was The Only Thing Connecting Claude, ChatGPT, and Codex. So I Built My Replacement.

By the end of this video, you're going By the end of this video, you're going to know how to make Claude, Codeex, to know how to make Claude, Codeex, to know how to make Claude, Codeex, Chad, GPT, and OpenClaw or Hermes work Chad, GPT, and OpenClaw or Hermes work Chad, GPT, and OpenClaw or Hermes work together without waiting for any of them together without waiting for any of them together without waiting for any of them to start to integrate with each other. to start to integrate with each other. to start to integrate with each other. And yes, I'm going to show demos. I'm And yes, I'm going to show demos. I'm And yes, I'm going to show demos. I'm going to show what I built. I'm going to going to show what I built. I'm going to going to show what I built. I'm going to talk to you about why. I'm going to give talk to you about why. I'm going to give talk to you about why. I'm going to give you real stories. I'm going to tell you you real stories. I'm going to tell you you real stories. I'm going to tell you I'm using it at home and at work. It's I'm using it at home and at work. It's I'm using it at home and at work. It's the full shebang. You're going to get the full shebang. You're going to get the full shebang. You're going to get the full tour. And by the end, you're the full tour. And by the end, you're the full tour. And by the end, you're going to be able to build it for going to be able to build it for going to be able to build it for yourself. I call it open engine. And the yourself. I call it open engine. And the yourself. I call it open engine. And the promise is simple. We need an open promise is simple. We need an open promise is simple. We need an open engine that drives our life because we engine that drives our life because we engine that drives our life because we have too many AIs and they don't talk have too many AIs and they don't talk have too many AIs and they don't talk together well. So, open engine gets your together well. So, open engine gets your together well. So, open engine gets your agents to stop acting like separate agents to stop acting like separate agents to stop acting like separate subscriptions or separate products and subscriptions or separate products and subscriptions or separate products and start acting like a system you can start acting like a system you can start acting like a system you can operate. And if you're wondering, is operate. And if you're wondering, is operate. And if you're wondering, is this for teams too? Yes, I'm using it this for teams too? Yes, I'm using it this for teams too? Yes, I'm using it with my team. It absolutely works for with my team. It absolutely works for with my team. It absolutely works for teams. It works for teams of humans and teams. It works for teams of humans and teams. It works for teams of humans and their agents and it's a seamless way to their agents and it's a seamless way to their agents and it's a seamless way to get them all to work together. I'm very get them all to work together. I'm very get them all to work together. I'm very excited. I'm going to show it off to you excited. I'm going to show it off to you excited. I'm going to show it off to you in this video. And yes, I've been in this video. And yes, I've been in this video. And yes, I've been building and using a working version of building and using a working version of building and using a working version of this to help me actually get stories this to help me actually get stories this to help me actually get stories out, organize my life, move houses, and out, organize my life, move houses, and out, organize my life, move houses, and I wanted to release it into the world I wanted to release it into the world I wanted to release it into the world because it's been so useful for me and because it's been so useful for me and because it's been so useful for me and it's actually lifted the load for my it's actually lifted the load for my it's actually lifted the load for my wife. Let me make this concrete with a wife. Let me make this concrete with a wife. Let me make this concrete with a real story of a friend of mine. She has real story of a friend of mine. She has real story of a friend of mine. She has a baby. She runs an agency. She uses a baby. She runs an agency. She uses a baby. She runs an agency. She uses Claude code. She's got loops and Claude code. She's got loops and Claude code. She's got loops and automations. She's looked seriously at automations. She's looked seriously at automations. She's looked seriously at OpenClaw because she wants agents that OpenClaw because she wants agents that OpenClaw because she wants agents that do real work. She is not trying AI for do real work. She is not trying AI for do real work. She is not trying AI for the first time. She's already using the the first time. She's already using the the first time. She's already using the tools and she's talking to me about tools and she's talking to me about tools and she's talking to me about them, right? Her challenge is that there them, right? Her challenge is that there them, right? Her challenge is that there are five at least different AI systems are five at least different AI systems are five at least different AI systems she's using that all help with a she's using that all help with a she's using that all help with a particular piece of the day and she particular piece of the day and she particular piece of the day and she becomes the person that carries the work becomes the person that carries the work becomes the person that carries the work between them. And that's a lot of labor

between them. And that's a lot of labor between them. And that's a lot of labor to carry. And anyone who's used these to carry. And anyone who's used these to carry. And anyone who's used these systems knows that that labor is real systems knows that that labor is real systems knows that that labor is real because you can't trade them out. Claude because you can't trade them out. Claude because you can't trade them out. Claude Code and Codeex don't do the same things Code and Codeex don't do the same things Code and Codeex don't do the same things even though they're aimed at the same even though they're aimed at the same even though they're aimed at the same segment of the population as a user segment of the population as a user segment of the population as a user base. Claude is better at front-end base. Claude is better at front-end base. Claude is better at front-end design. It just is. It's intuitive. design. It just is. It's intuitive. design. It just is. It's intuitive. OpenAI is less good at that, but OpenAI OpenAI is less good at that, but OpenAI OpenAI is less good at that, but OpenAI has a reputation for back-end has a reputation for back-end has a reputation for back-end engineering that's very strong. And a engineering that's very strong. And a engineering that's very strong. And a lot of us who understand these things lot of us who understand these things lot of us who understand these things are juggling a lot and it's painful and are juggling a lot and it's painful and are juggling a lot and it's painful and we've had to effectively drive our own we've had to effectively drive our own we've had to effectively drive our own harnesses to make up for that by harnesses to make up for that by harnesses to make up for that by coordinating ourselves across all of coordinating ourselves across all of coordinating ourselves across all of these tools. So, she's using five these tools. So, she's using five these tools. So, she's using five different AI tools, and the question is different AI tools, and the question is different AI tools, and the question is how work can leave a tool, land with the how work can leave a tool, land with the how work can leave a tool, land with the right person or the right next agent, right person or the right next agent, right person or the right next agent, bring the source material with it, show bring the source material with it, show bring the source material with it, show what happened, and not make anybody read what happened, and not make anybody read what happened, and not make anybody read a giant chat transcript. And if that a giant chat transcript. And if that a giant chat transcript. And if that sounds like you, I've got good news for sounds like you, I've got good news for sounds like you, I've got good news for you. I'm putting together an agent tool you. I'm putting together an agent tool you. I'm putting together an agent tool that solves for that. And that's what that solves for that. And that's what that solves for that. And that's what Open Engine really is. It is a tool that Open Engine really is. It is a tool that Open Engine really is. It is a tool that allows every AI in your system to allows every AI in your system to allows every AI in your system to coordinate seamlessly and carry state or coordinate seamlessly and carry state or coordinate seamlessly and carry state or context or detail back and forth without context or detail back and forth without context or detail back and forth without you having to do the work. She is trying you having to do the work. She is trying you having to do the work. She is trying to figure out how to balance a client to figure out how to balance a client to figure out how to balance a client call, a product scoping conversation, call, a product scoping conversation, call, a product scoping conversation, and a baby appointment that just came and a baby appointment that just came and a baby appointment that just came up. She has to figure out how to do all up. She has to figure out how to do all up. She has to figure out how to do all three of those. She typically uses three of those. She typically uses three of those. She typically uses codeex for product scoping. She's using codeex for product scoping. She's using codeex for product scoping. She's using clawed code to reorganize and move her clawed code to reorganize and move her clawed code to reorganize and move her calendar around and she's trying to deal calendar around and she's trying to deal calendar around and she's trying to deal with the baby appointment by email, but

with the baby appointment by email, but with the baby appointment by email, but she would like to find some automation she would like to find some automation she would like to find some automation there for that. There is no easy way to there for that. There is no easy way to there for that. There is no easy way to tackle all of that in one thing unless tackle all of that in one thing unless tackle all of that in one thing unless you're compromising on models somewhere. you're compromising on models somewhere. you're compromising on models somewhere. And so the question that she has is can And so the question that she has is can And so the question that she has is can she live without compromises? Can she she live without compromises? Can she she live without compromises? Can she find a way to get her preferred model find a way to get her preferred model find a way to get her preferred model against a particular problem and not against a particular problem and not against a particular problem and not feel like she has to trade that off in feel like she has to trade that off in feel like she has to trade that off in order to keep track of everything? And order to keep track of everything? And order to keep track of everything? And kind of going farther than that, can she kind of going farther than that, can she kind of going farther than that, can she avoid having to depend on unpredictable avoid having to depend on unpredictable avoid having to depend on unpredictable memory in order to do that? And one of memory in order to do that? And one of memory in order to do that? And one of the challenges with OpenClaw, and I've the challenges with OpenClaw, and I've the challenges with OpenClaw, and I've installed it, I've used it, is that it's installed it, I've used it, is that it's installed it, I've used it, is that it's sometimes not entirely accurate when sometimes not entirely accurate when sometimes not entirely accurate when you're doing multiple different roles in you're doing multiple different roles in you're doing multiple different roles in your life over a long enough period of your life over a long enough period of your life over a long enough period of time. And so in this case, like when you time. And so in this case, like when you time. And so in this case, like when you have the the baby appointment pop back have the the baby appointment pop back have the the baby appointment pop back up after a couple of months and then at up after a couple of months and then at up after a couple of months and then at the same time, you have an agency team the same time, you have an agency team the same time, you have an agency team question and your open claw can't talk question and your open claw can't talk question and your open claw can't talk to the team intuitively unless you give to the team intuitively unless you give to the team intuitively unless you give it permissions on Slack. And then if you it permissions on Slack. And then if you it permissions on Slack. And then if you do, it doesn't talk to the team's agent. do, it doesn't talk to the team's agent. do, it doesn't talk to the team's agent. And then at the same time you have to And then at the same time you have to And then at the same time you have to get it into the email somehow and then get it into the email somehow and then get it into the email somehow and then there's a whole memory piece that goes there's a whole memory piece that goes there's a whole memory piece that goes this is the actual conversation I hear.

this is the actual conversation I hear. this is the actual conversation I hear. Are you hearing how many ifs, buts, Are you hearing how many ifs, buts, Are you hearing how many ifs, buts, wins, ifs, and copies and paste there wins, ifs, and copies and paste there wins, ifs, and copies and paste there are there's so much it's a lot of are there's so much it's a lot of are there's so much it's a lot of extended work that we're carrying for extended work that we're carrying for extended work that we're carrying for using these AI tools and people who are using these AI tools and people who are using these AI tools and people who are AI productive are just carrying that AI productive are just carrying that AI productive are just carrying that load because the AI payoff is so great. load because the AI payoff is so great. load because the AI payoff is so great. And what we're seeing in 2026 is that AI And what we're seeing in 2026 is that AI And what we're seeing in 2026 is that AI is helping with these pieces. And post is helping with these pieces. And post is helping with these pieces. And post openclaw, we're getting some of that openclaw, we're getting some of that openclaw, we're getting some of that coordination piece in, but the really coordination piece in, but the really coordination piece in, but the really hard part isn't solved yet. The really hard part isn't solved yet. The really hard part isn't solved yet. The really hard part is the movement between the hard part is the movement between the hard part is the movement between the pieces seamlessly and carrying full pieces seamlessly and carrying full pieces seamlessly and carrying full state, all the details, right? So on state, all the details, right? So on state, all the details, right? So on Wednesday, I talked about the idea that Wednesday, I talked about the idea that Wednesday, I talked about the idea that agents are really loop managers. A agents are really loop managers. A agents are really loop managers. A useful agent is a remembered workflow useful agent is a remembered workflow useful agent is a remembered workflow that can run again and notice what that can run again and notice what that can run again and notice what changed and stop in the right place and changed and stop in the right place and changed and stop in the right place and bring you in when the decision is really bring you in when the decision is really bring you in when the decision is really needed. And that is absolutely the right needed. And that is absolutely the right needed. And that is absolutely the right basic frame for an agent. But if every basic frame for an agent. But if every basic frame for an agent. But if every loop lives in its own room, the human loop lives in its own room, the human loop lives in its own room, the human becomes the hallway. The research loop becomes the hallway. The research loop becomes the hallway. The research loop finishes and the writing loop doesn't finishes and the writing loop doesn't finishes and the writing loop doesn't know what changed. The support loop sees know what changed. The support loop sees know what changed. The support loop sees a pattern, but the product loop doesn't a pattern, but the product loop doesn't a pattern, but the product loop doesn't get the original messages unless you put get the original messages unless you put get the original messages unless you put them there. Right? The coding agent them there. Right? The coding agent them there. Right? The coding agent fixes the file, but the teammate who fixes the file, but the teammate who fixes the file, but the teammate who owns review only sees a vague summary owns review only sees a vague summary owns review only sees a vague summary unless you send them a bunch of chats.

unless you send them a bunch of chats. unless you send them a bunch of chats. The schedule changes, but the work loop The schedule changes, but the work loop The schedule changes, but the work loop doesn't know that the afternoon just doesn't know that the afternoon just doesn't know that the afternoon just collapsed. Going back to our story, this collapsed. Going back to our story, this collapsed. Going back to our story, this is why I built Open Engine. And this is is why I built Open Engine. And this is is why I built Open Engine. And this is the kind of problem I've been working on the kind of problem I've been working on the kind of problem I've been working on solving for a while. So if you recall, solving for a while. So if you recall, solving for a while. So if you recall, Open Brain was about memory. So a few Open Brain was about memory. So a few Open Brain was about memory. So a few months ago, I argued that every AI you months ago, I argued that every AI you months ago, I argued that every AI you use starts from zero unless you give it use starts from zero unless you give it use starts from zero unless you give it a true memory system that you control a true memory system that you control a true memory system that you control and that lives between your agents. Your and that lives between your agents. Your and that lives between your agents. Your context should not be trapped inside one context should not be trapped inside one context should not be trapped inside one company's chat history. Open engine is company's chat history. Open engine is company's chat history. Open engine is the next big piece here. Once the AI can the next big piece here. Once the AI can the next big piece here. Once the AI can remember how does your work actually remember how does your work actually remember how does your work actually move and and the basic approach here is move and and the basic approach here is move and and the basic approach here is very very simple and I'm doing this and very very simple and I'm doing this and very very simple and I'm doing this and talking about it as simply as possible talking about it as simply as possible talking about it as simply as possible because I want this to be easy for you. because I want this to be easy for you. because I want this to be easy for you. Open engine is the next missing piece in Open engine is the next missing piece in Open engine is the next missing piece in the story for all of us who are trying the story for all of us who are trying the story for all of us who are trying to be productive with AI and actually to be productive with AI and actually to be productive with AI and actually get rid of all of that invisible work. get rid of all of that invisible work. get rid of all of that invisible work. Once the AI can remember how does our Once the AI can remember how does our Once the AI can remember how does our work move so we don't have to spend a work move so we don't have to spend a work move so we don't have to spend a lot of time coordinating and doing that lot of time coordinating and doing that lot of time coordinating and doing that invisible labor. The basic move that I'm invisible labor. The basic move that I'm invisible labor. The basic move that I'm going to propose here is extremely going to propose here is extremely going to propose here is extremely simple on purpose. I want to make this simple on purpose. I want to make this simple on purpose. I want to make this as easy to use and do as possible. Just as easy to use and do as possible. Just as easy to use and do as possible. Just put the work in a queue that both people put the work in a queue that both people put the work in a queue that both people and agents can read. What is a queue? It and agents can read. What is a queue? It and agents can read. What is a queue? It can be as simple as a Jira system. It can be as simple as a Jira system. It can be as simple as a Jira system. It can be a a conbon board that you coded can be a a conbon board that you coded can be a a conbon board that you coded up. It can be a linear ticket Q. That's up. It can be a linear ticket Q. That's up. It can be a linear ticket Q. That's what I like. Whatever it is, as long as what I like. Whatever it is, as long as what I like. Whatever it is, as long as it is a Q that an agent can write to and it is a Q that an agent can write to and it is a Q that an agent can write to and you can read to, it's good enough. And

you can read to, it's good enough. And you can read to, it's good enough. And if you have a queue like that, then all if you have a queue like that, then all if you have a queue like that, then all you need to do is have an issue that you need to do is have an issue that you need to do is have an issue that says this is what needs to happen. This says this is what needs to happen. This says this is what needs to happen. This is who owns it. this is the background is who owns it. this is the background is who owns it. this is the background that matters. This is what the agent can that matters. This is what the agent can that matters. This is what the agent can do and where the agent should stop and do and where the agent should stop and do and where the agent should stop and what it has to show when it's done. And what it has to show when it's done. And what it has to show when it's done. And that sounds really, really simple. And that sounds really, really simple. And that sounds really, really simple. And simple is the point, right? A good Q simple is the point, right? A good Q simple is the point, right? A good Q issue or a good ticket is the difference issue or a good ticket is the difference issue or a good ticket is the difference between asking an AI for help and giving between asking an AI for help and giving between asking an AI for help and giving it a job the next person, the next agent it a job the next person, the next agent it a job the next person, the next agent can understand so it actually gets off can understand so it actually gets off can understand so it actually gets off your plate and gets done. Right? If we your plate and gets done. Right? If we your plate and gets done. Right? If we go back to 2025 when we were talking a go back to 2025 when we were talking a go back to 2025 when we were talking a lot about prompting, a prompt asks for lot about prompting, a prompt asks for lot about prompting, a prompt asks for an answer of some sort. A ticket asks an answer of some sort. A ticket asks an answer of some sort. A ticket asks for a result to get done and it can have for a result to get done and it can have for a result to get done and it can have multiple agents even if the agents don't multiple agents even if the agents don't multiple agents even if the agents don't know each other and aren't directly know each other and aren't directly know each other and aren't directly integrated. The ticket becomes the place integrated. The ticket becomes the place integrated. The ticket becomes the place they talk. And I think that this they talk. And I think that this they talk. And I think that this distinction is really important because distinction is really important because distinction is really important because as agents get better, we need better as agents get better, we need better as agents get better, we need better state management for our agents. A chat state management for our agents. A chat state management for our agents. A chat box is a terrible way to manage state.

box is a terrible way to manage state. box is a terrible way to manage state. And I'm sorry, but so is Slack. An agent And I'm sorry, but so is Slack. An agent And I'm sorry, but so is Slack. An agent needs to be able to change files and needs to be able to change files and needs to be able to change files and create tasks and move statuses and write create tasks and move statuses and write create tasks and move statuses and write drafts and do all of that in a place drafts and do all of that in a place drafts and do all of that in a place where you can actually see what where you can actually see what where you can actually see what happened. When we're picking paint happened. When we're picking paint happened. When we're picking paint colors for the house, I need to see what colors for the house, I need to see what colors for the house, I need to see what kind of paint we picked out, right? Like kind of paint we picked out, right? Like kind of paint we picked out, right? Like otherwise, we're going to get all mixed otherwise, we're going to get all mixed otherwise, we're going to get all mixed up and it just lives in somebody's head up and it just lives in somebody's head up and it just lives in somebody's head and then we're doing that mental labor. and then we're doing that mental labor. and then we're doing that mental labor. When my friend is trying to figure out When my friend is trying to figure out When my friend is trying to figure out whether she has five or six different whether she has five or six different whether she has five or six different companies in pipeline, she actually has companies in pipeline, she actually has companies in pipeline, she actually has to be able to audit that pipeline from to be able to audit that pipeline from to be able to audit that pipeline from any given command line that she has. See any given command line that she has. See any given command line that she has. See what has been done on it by her team, what has been done on it by her team, what has been done on it by her team, what has been done on it by her team's what has been done on it by her team's what has been done on it by her team's agents, what has been done on it by her agents, what has been done on it by her agents, what has been done on it by her agents, and what the next step is. agents, and what the next step is. agents, and what the next step is. Again, you can't do that by querying Again, you can't do that by querying Again, you can't do that by querying chats. So, this is what I'm trying to chats. So, this is what I'm trying to chats. So, this is what I'm trying to build with Open Engine. And and the build with Open Engine. And and the build with Open Engine. And and the guide that I put together is very guide that I put together is very guide that I put together is very specific. It has five different specific. It has five different specific. It has five different components that you can point at that components that you can point at that components that you can point at that all add up together into a complete all add up together into a complete all add up together into a complete ecosystem. And if you miss one of them, ecosystem. And if you miss one of them, ecosystem. And if you miss one of them, it doesn't really go together, right? It it doesn't really go together, right? It it doesn't really go together, right? It doesn't hang together. The first is a doesn't hang together. The first is a doesn't hang together. The first is a linear cube. That makes a lot of sense.

linear cube. That makes a lot of sense. linear cube. That makes a lot of sense. Linear's got a generous free plan. It Linear's got a generous free plan. It Linear's got a generous free plan. It works well with with all of the AI works well with with all of the AI works well with with all of the AI systems. That's why I chose it. But you systems. That's why I chose it. But you systems. That's why I chose it. But you don't have to use L. You can use Jura. don't have to use L. You can use Jura. don't have to use L. You can use Jura. You can use your own system. It's fun. You can use your own system. It's fun. You can use your own system. It's fun. Then we get into how you tell your AI, Then we get into how you tell your AI, Then we get into how you tell your AI, hey, this is how you use this tool. hey, this is how you use this tool. hey, this is how you use this tool. Because the AI needs to be told, this is Because the AI needs to be told, this is Because the AI needs to be told, this is the protocol for using this ticketing the protocol for using this ticketing the protocol for using this ticketing system. And so there are four other system. And so there are four other system. And so there are four other pieces that go with that. And I've pieces that go with that. And I've pieces that go with that. And I've written basically skills that tell your written basically skills that tell your written basically skills that tell your AI how to use this, right? There's a AI how to use this, right? There's a AI how to use this, right? There's a setup skill, there's a status skill, setup skill, there's a status skill, setup skill, there's a status skill, there's there's a skill to run a a cue there's there's a skill to run a a cue there's there's a skill to run a a cue through the the AI, and then there's a through the the AI, and then there's a through the the AI, and then there's a smoke test. You can start to test this. smoke test. You can start to test this. smoke test. You can start to test this. I want this actually to help me. And I want this actually to help me. And I want this actually to help me. And yes, you can absolutely point your yes, you can absolutely point your yes, you can absolutely point your openclaw or your Hermes at this as a openclaw or your Hermes at this as a openclaw or your Hermes at this as a skill and use that too, right? This is skill and use that too, right? This is skill and use that too, right? This is not an anti-openclaw sentiment. I love not an anti-openclaw sentiment. I love not an anti-openclaw sentiment. I love what Open Claw has done. I love what what Open Claw has done. I love what what Open Claw has done. I love what Peter's done. I love Hermes. You can use Peter's done. I love Hermes. You can use Peter's done. I love Hermes. You can use those tools as much as you want and those tools as much as you want and those tools as much as you want and you're not limited by only using those you're not limited by only using those you're not limited by only using those tools. Open Claw is pointed at a real tools. Open Claw is pointed at a real tools. Open Claw is pointed at a real desire that made a lot of sense. We want desire that made a lot of sense. We want desire that made a lot of sense. We want agents that can act, right? Agents that agents that can act, right? Agents that agents that can act, right? Agents that aren't a chat window. Hermes and similar aren't a chat window. Hermes and similar aren't a chat window. Hermes and similar projects are aimed at another real projects are aimed at another real projects are aimed at another real desire that goes beyond that. Agents can desire that goes beyond that. Agents can desire that goes beyond that. Agents can improve at repeated work and learn improve at repeated work and learn improve at repeated work and learn instead of starting cold every time.

instead of starting cold every time. instead of starting cold every time. These are real needs. But making an These are real needs. But making an These are real needs. But making an agent autonomous is not enough. A very agent autonomous is not enough. A very agent autonomous is not enough. A very capable private agent can still become capable private agent can still become capable private agent can still become another inbox or a task cue or a text another inbox or a task cue or a text another inbox or a task cue or a text message cue that you have to manage. The message cue that you have to manage. The message cue that you have to manage. The bottleneck I care about is the boundary bottleneck I care about is the boundary bottleneck I care about is the boundary between agents because that's where I between agents because that's where I between agents because that's where I feel the pain. Can the work leave Claude feel the pain. Can the work leave Claude feel the pain. Can the work leave Claude and go to codeex? Can a teammate's agent and go to codeex? Can a teammate's agent and go to codeex? Can a teammate's agent pick up a task created by my agent? Can pick up a task created by my agent? Can pick up a task created by my agent? Can a support loop escalate to the person a support loop escalate to the person a support loop escalate to the person with authority without losing any of the with authority without losing any of the with authority without losing any of the message or customer history and the message or customer history and the message or customer history and the reason for the agent stopping work? All reason for the agent stopping work? All reason for the agent stopping work? All of this stuff is not a model problem. of this stuff is not a model problem. of this stuff is not a model problem. It's not an agent problem. It's a It's not an agent problem. It's a It's not an agent problem. It's a boundary problem. It's a who gets this boundary problem. It's a who gets this boundary problem. It's a who gets this next and how do we hand it off problem. next and how do we hand it off problem. next and how do we hand it off problem. If I go back to my friend, her problem If I go back to my friend, her problem If I go back to my friend, her problem is actually much more painful than a is actually much more painful than a is actually much more painful than a tool problem. It's much more painful tool problem. It's much more painful tool problem. It's much more painful than a fluency problem. She's fluent in than a fluency problem. She's fluent in than a fluency problem. She's fluent in AI tools. That's not an issue. She's got AI tools. That's not an issue. She's got AI tools. That's not an issue. She's got lots of AI tools. That's not an issue. lots of AI tools. That's not an issue. lots of AI tools. That's not an issue. The problem is a handoff problem. The The problem is a handoff problem. The The problem is a handoff problem. The problem is the information isn't flowing problem is the information isn't flowing problem is the information isn't flowing between her, her agents, her team's between her, her agents, her team's between her, her agents, her team's agents, and her team smoothly, fluently, agents, and her team smoothly, fluently, agents, and her team smoothly, fluently, without someone having to manage it. And without someone having to manage it. And without someone having to manage it. And it would, it is amazing to me how much it would, it is amazing to me how much it would, it is amazing to me how much of AI promise gets bogged down in those of AI promise gets bogged down in those of AI promise gets bogged down in those handoff points. If you look at AI as handoff points. If you look at AI as handoff points. If you look at AI as essentially a technology change that essentially a technology change that essentially a technology change that moves bottlenecks, we are moving a ton moves bottlenecks, we are moving a ton moves bottlenecks, we are moving a ton of generative energy into this tiny of generative energy into this tiny of generative energy into this tiny bottleneck around handoffs. And that is bottleneck around handoffs. And that is bottleneck around handoffs. And that is what Open Engine is designed to attack what Open Engine is designed to attack what Open Engine is designed to attack and change and blow open so it gets

and change and blow open so it gets and change and blow open so it gets easier. And yes, this works for easier. And yes, this works for easier. And yes, this works for households and teams at the same time. households and teams at the same time. households and teams at the same time. You think about it, a school pickup You think about it, a school pickup You think about it, a school pickup change is not the same as a sales change is not the same as a sales change is not the same as a sales pipeline, but the shape of handoff pain pipeline, but the shape of handoff pain pipeline, but the shape of handoff pain underneath is very familiar. Something underneath is very familiar. Something underneath is very familiar. Something has changed. Several other things now has changed. Several other things now has changed. Several other things now depend on it. Some parts can be handled depend on it. Some parts can be handled depend on it. Some parts can be handled by agents. Some parts need a person. Uh, by agents. Some parts need a person. Uh, by agents. Some parts need a person. Uh, and Open Engine doesn't replace the and Open Engine doesn't replace the and Open Engine doesn't replace the person. All it does is keep the person person. All it does is keep the person person. All it does is keep the person from being the only one handing all of from being the only one handing all of from being the only one handing all of this stuff off. And that is such a huge this stuff off. And that is such a huge this stuff off. And that is such a huge load off our shoulders. We found it's a load off our shoulders. We found it's a load off our shoulders. We found it's a huge load off for the family. And I huge load off for the family. And I huge load off for the family. And I found working with my team, it's a found working with my team, it's a found working with my team, it's a massive load off of our shoulders as massive load off of our shoulders as massive load off of our shoulders as well because our agents can now work well because our agents can now work well because our agents can now work together seamlessly. And we're not together seamlessly. And we're not together seamlessly. And we're not trying to track agent messages in Slack trying to track agent messages in Slack trying to track agent messages in Slack necessarily, although we can invoke our necessarily, although we can invoke our necessarily, although we can invoke our agents through Slack. We actually have a agents through Slack. We actually have a agents through Slack. We actually have a clear system of record and we can tag clear system of record and we can tag clear system of record and we can tag each other's agents in and it becomes a each other's agents in and it becomes a each other's agents in and it becomes a really effective way to collaborate really effective way to collaborate really effective way to collaborate without losing track of any detail of without losing track of any detail of without losing track of any detail of what we're doing. Think about it. If the what we're doing. Think about it. If the what we're doing. Think about it. If the agent writes a beautiful brief in a agent writes a beautiful brief in a agent writes a beautiful brief in a private chat and nobody knows where it private chat and nobody knows where it private chat and nobody knows where it came from, it's just a draft in a room came from, it's just a draft in a room came from, it's just a draft in a room by itself. Output is what the AI returns by itself. Output is what the AI returns by itself. Output is what the AI returns right now. Work is what someone can right now. Work is what someone can right now. Work is what someone can review, accept, and build on. And open review, accept, and build on. And open review, accept, and build on. And open engine is about getting from output to engine is about getting from output to engine is about getting from output to work without making human beings the work without making human beings the work without making human beings the copy paste path. So jumping into demo, copy paste path. So jumping into demo, copy paste path. So jumping into demo, this is the basic open engine loop that this is the basic open engine loop that this is the basic open engine loop that you see on the screen. It's a request you see on the screen. It's a request you see on the screen. It's a request that becomes a clear linear task. It's that becomes a clear linear task. It's that becomes a clear linear task. It's assigned to the right operator. Codeex assigned to the right operator. Codeex assigned to the right operator. Codeex wakes up. It checks its assigned queue wakes up. It checks its assigned queue wakes up. It checks its assigned queue and finds one eligible agent and finds one eligible agent and finds one eligible agent instructions issue. Before doing the

instructions issue. Before doing the instructions issue. Before doing the work, it claim locks the issue. It moves work, it claim locks the issue. It moves work, it claim locks the issue. It moves it to agent working and it leaves an it to agent working and it leaves an it to agent working and it leaves an agent claimed receipt. And then agent claimed receipt. And then agent claimed receipt. And then execution starts locally, right? Linear execution starts locally, right? Linear execution starts locally, right? Linear coordinates the team. Codex does the coordinates the team. Codex does the coordinates the team. Codex does the work. A human can create a task for that work. A human can create a task for that work. A human can create a task for that agent and an agent create a task for agent and an agent create a task for agent and an agent create a task for another agent. The task includes the another agent. The task includes the another agent. The task includes the outcome and sources and the definition outcome and sources and the definition outcome and sources and the definition of done. And when the agent starts, it of done. And when the agent starts, it of done. And when the agent starts, it moves from agent to-do to agent working. moves from agent to-do to agent working. moves from agent to-do to agent working. And when it finishes, it also moves the And when it finishes, it also moves the And when it finishes, it also moves the issue to the right place. The receipt is issue to the right place. The receipt is issue to the right place. The receipt is not decoration here. It actually lets not decoration here. It actually lets not decoration here. It actually lets you know what was done. I don't want to you know what was done. I don't want to you know what was done. I don't want to ask the agent, did you do the thing? I ask the agent, did you do the thing? I ask the agent, did you do the thing? I want to know it got done. And I don't want to know it got done. And I don't want to know it got done. And I don't want to have to copy and paste. I don't want to have to copy and paste. I don't want to have to copy and paste. I don't want to have to coordinate between my my want to have to coordinate between my my want to have to coordinate between my my LLMs anymore. I'm so tired of it. So, LLMs anymore. I'm so tired of it. So, LLMs anymore. I'm so tired of it. So, this is how you actually take all of the this is how you actually take all of the this is how you actually take all of the context, all of the artifacts, and move context, all of the artifacts, and move context, all of the artifacts, and move it into a system where agents can it into a system where agents can it into a system where agents can actually tackle that work in a actually tackle that work in a actually tackle that work in a transparent manner. So, humans don't get transparent manner. So, humans don't get transparent manner. So, humans don't get confused about what's being worked on. confused about what's being worked on. confused about what's being worked on. So, agents don't get confused. It just So, agents don't get confused. It just So, agents don't get confused. It just becomes very simple. All right, let's becomes very simple. All right, let's becomes very simple. All right, let's take a look at delegation, which is a take a look at delegation, which is a take a look at delegation, which is a key part of this pattern. Can you key part of this pattern. Can you key part of this pattern. Can you delegate to agents? And what does that delegate to agents? And what does that delegate to agents? And what does that look like? Let's say Maya asks Codex to look like? Let's say Maya asks Codex to look like? Let's say Maya asks Codex to route a metric sp to Leo's agent. Leo's route a metric sp to Leo's agent. Leo's route a metric sp to Leo's agent. Leo's agent happens to be Claude. Codex checks agent happens to be Claude. Codex checks agent happens to be Claude. Codex checks that Leo's agent is online and then that Leo's agent is online and then that Leo's agent is online and then writes a self-contained linear issue writes a self-contained linear issue writes a self-contained linear issue with the context needed to act. That's with the context needed to act. That's with the context needed to act. That's two people's different agents from two people's different agents from two people's different agents from different LLM providers coordinating and different LLM providers coordinating and different LLM providers coordinating and linear. It assigns the issue to Leo's linear. It assigns the issue to Leo's linear. It assigns the issue to Leo's agent, keeps it in agent to-do, and agent, keeps it in agent to-do, and agent, keeps it in agent to-do, and labels it as agent instructions. Now, labels it as agent instructions. Now, labels it as agent instructions. Now, Leo's agent can pick it up on its own

Leo's agent can pick it up on its own Leo's agent can pick it up on its own heartbeat and the handoff is visible and heartbeat and the handoff is visible and heartbeat and the handoff is visible and it's scoped. It's not buried in chat it's scoped. It's not buried in chat it's scoped. It's not buried in chat because the point is not that we made an because the point is not that we made an because the point is not that we made an agent do a trick here, right? The point agent do a trick here, right? The point agent do a trick here, right? The point is you can actually see the work move is you can actually see the work move is you can actually see the work move through the team. And in that context, through the team. And in that context, through the team. And in that context, the smoke test in the skill, right? It's the smoke test in the skill, right? It's the smoke test in the skill, right? It's deliberately clean and very, very deliberately clean and very, very deliberately clean and very, very efficient. All you're doing is saying efficient. All you're doing is saying efficient. All you're doing is saying create an issue called say hello from create an issue called say hello from create an issue called say hello from the queue. You're assigning it to an the queue. You're assigning it to an the queue. You're assigning it to an human or an agent, giving it an human or an agent, giving it an human or an agent, giving it an instruction label, and making sure that instruction label, and making sure that instruction label, and making sure that it actually can move to done. So, this it actually can move to done. So, this it actually can move to done. So, this is not about giving your agent a bunch is not about giving your agent a bunch is not about giving your agent a bunch of work from me. This is just about of work from me. This is just about of work from me. This is just about giving you a clean smoke test so you can giving you a clean smoke test so you can giving you a clean smoke test so you can see that this agent interaction loop see that this agent interaction loop see that this agent interaction loop works. The full loop is actually pretty works. The full loop is actually pretty works. The full loop is actually pretty simple. You have to have a request. You simple. You have to have a request. You simple. You have to have a request. You have to have an issue, a claim, a piece have to have an issue, a claim, a piece have to have an issue, a claim, a piece of work that's done, some proof that of work that's done, some proof that of work that's done, some proof that it's done, a receipt, and then you want it's done, a receipt, and then you want it's done, a receipt, and then you want to go on to the next item. Let's say Leo to go on to the next item. Let's say Leo to go on to the next item. Let's say Leo the agent claims the task and moves it the agent claims the task and moves it the agent claims the task and moves it into agent working so everyone knows into agent working so everyone knows into agent working so everyone knows it's worked on. If it hits ambiguity, it's worked on. If it hits ambiguity, it's worked on. If it hits ambiguity, the agent doesn't guess. Instead, it the agent doesn't guess. Instead, it the agent doesn't guess. Instead, it moves to needs input and asks the exact moves to needs input and asks the exact moves to needs input and asks the exact blocking question. Maya the human can blocking question. Maya the human can blocking question. Maya the human can answer on that issue and the agent can answer on that issue and the agent can answer on that issue and the agent can resume and the audit trail stays in one resume and the audit trail stays in one resume and the audit trail stays in one place. And when finished, Leo leaves place. And when finished, Leo leaves place. And when finished, Leo leaves agent done as a status and moves on to agent done as a status and moves on to agent done as a status and moves on to the next task. And we can use that for the next task. And we can use that for the next task. And we can use that for any piece of work. You can use that for any piece of work. You can use that for any piece of work. You can use that for picking paint colors in the house as picking paint colors in the house as picking paint colors in the house as easily as you can use it for scheduling easily as you can use it for scheduling easily as you can use it for scheduling a pipeline review while you are on the a pipeline review while you are on the a pipeline review while you are on the go trying to sort out your kids's doctor go trying to sort out your kids's doctor go trying to sort out your kids's doctor appointment. The the point is that the appointment. The the point is that the appointment. The the point is that the queue itself makes it easy to sort out

queue itself makes it easy to sort out queue itself makes it easy to sort out all of the handoff stress so that you're all of the handoff stress so that you're all of the handoff stress so that you're not dealing with it because open engine not dealing with it because open engine not dealing with it because open engine demands that we move from prompt mode to demands that we move from prompt mode to demands that we move from prompt mode to demanding real work. And the framework demanding real work. And the framework demanding real work. And the framework helps us get there. Open engine helps us helps us get there. Open engine helps us helps us get there. Open engine helps us get there. So, prompt mode might be, get there. So, prompt mode might be, get there. So, prompt mode might be, "Write me a follow-up email." Work mode "Write me a follow-up email." Work mode "Write me a follow-up email." Work mode might be, "Here's the client call might be, "Here's the client call might be, "Here's the client call transcript. Here's the decision we made. transcript. Here's the decision we made. transcript. Here's the decision we made. Here's the promise. I don't want to Here's the promise. I don't want to Here's the promise. I don't want to overstate. Here's the calendar overstate. Here's the calendar overstate. Here's the calendar constraints. You draft the follow-up constraints. You draft the follow-up constraints. You draft the follow-up agent, flag what needs my judgment, and agent, flag what needs my judgment, and agent, flag what needs my judgment, and leave notes that I can review later." leave notes that I can review later." leave notes that I can review later." See, that's a full statement of work See, that's a full statement of work See, that's a full statement of work right there. And you can actually pass right there. And you can actually pass right there. And you can actually pass that to another agent to review using that to another agent to review using that to another agent to review using Open Engine 2 if you want a second Open Engine 2 if you want a second Open Engine 2 if you want a second opinion, which a lot of people do, opinion, which a lot of people do, opinion, which a lot of people do, especially on client-f facing especially on client-f facing especially on client-f facing communication. You could have a whole communication. You could have a whole communication. You could have a whole agent that just reviews for brand agent that just reviews for brand agent that just reviews for brand language. I'll give you another couple language. I'll give you another couple language. I'll give you another couple examples. Prompt mode might be, hey, examples. Prompt mode might be, hey, examples. Prompt mode might be, hey, summarize this support ticket. You ask summarize this support ticket. You ask summarize this support ticket. You ask your codeex to do that. You ask your your codeex to do that. You ask your your codeex to do that. You ask your claw to do that. Work mode would be claw to do that. Work mode would be claw to do that. Work mode would be classify this ticket, attach the classify this ticket, attach the classify this ticket, attach the customer history, identify whether this customer history, identify whether this customer history, identify whether this was a known issue, create a product task was a known issue, create a product task was a known issue, create a product task only if it meets my escalation rule, and only if it meets my escalation rule, and only if it meets my escalation rule, and show exactly where you stopped work and show exactly where you stopped work and show exactly where you stopped work and why. Or this one, prompt mode would be, why. Or this one, prompt mode would be, why. Or this one, prompt mode would be, "Help me change my schedule." Work mode "Help me change my schedule." Work mode "Help me change my schedule." Work mode would be the pickup time for the kids would be the pickup time for the kids would be the pickup time for the kids has changed. Check what it affects.

has changed. Check what it affects. has changed. Check what it affects. draft the two messages that might be draft the two messages that might be draft the two messages that might be needed to the school and wait for needed to the school and wait for needed to the school and wait for approval before anything leaves the approval before anything leaves the approval before anything leaves the system. The model can be the same here, system. The model can be the same here, system. The model can be the same here, but the assignment is much more clear but the assignment is much more clear but the assignment is much more clear and the ability to hand off to multiple and the ability to hand off to multiple and the ability to hand off to multiple agents or to a human as you determine is agents or to a human as you determine is agents or to a human as you determine is really really easy. We need to stop really really easy. We need to stop really really easy. We need to stop assuming that we are the glue between assuming that we are the glue between assuming that we are the glue between all of the AI systems that we work for. all of the AI systems that we work for. all of the AI systems that we work for. Otherwise, the the AI is prompting us Otherwise, the the AI is prompting us Otherwise, the the AI is prompting us and we're just working for the AI and we're just working for the AI and we're just working for the AI instead of the other way around. We want instead of the other way around. We want instead of the other way around. We want to be in a place where the work can move to be in a place where the work can move to be in a place where the work can move across the tools that we're already across the tools that we're already across the tools that we're already using. And that's the open engine using. And that's the open engine using. And that's the open engine promise because our teams don't really promise because our teams don't really promise because our teams don't really live in one AI tool. I don't know of live in one AI tool. I don't know of live in one AI tool. I don't know of anyone who is curious about AI who anyone who is curious about AI who anyone who is curious about AI who actually uses only one tool. And actually uses only one tool. And actually uses only one tool. And especially if you get two people especially if you get two people especially if you get two people together, it's not one tool. It's like together, it's not one tool. It's like together, it's not one tool. It's like two different tools, three, five. I two different tools, three, five. I two different tools, three, five. I often see six different tools between often see six different tools between often see six different tools between two people. And you're not going to two people. And you're not going to two people. And you're not going to expect people to actually change that. expect people to actually change that. expect people to actually change that. you actually just want to give people a you actually just want to give people a you actually just want to give people a quue where all of their AIs can talk to quue where all of their AIs can talk to quue where all of their AIs can talk to each other. So, I'm not going to make each other. So, I'm not going to make each other. So, I'm not going to make you a big promise here. I'm not saying you a big promise here. I'm not saying you a big promise here. I'm not saying Open Engine can run your company. I Open Engine can run your company. I Open Engine can run your company. I don't want to say that. I don't want to don't want to say that. I don't want to don't want to say that. I don't want to say Open Engine can run your household.

say Open Engine can run your household. say Open Engine can run your household. What I want to say is the next useful AI What I want to say is the next useful AI What I want to say is the next useful AI is not another private assistant. It's a is not another private assistant. It's a is not another private assistant. It's a cue where your agents can all work cue where your agents can all work cue where your agents can all work together and move stuff forward. And together and move stuff forward. And together and move stuff forward. And that matters on any team. But that matters on any team. But that matters on any team. But increasingly, we're all managing teams increasingly, we're all managing teams increasingly, we're all managing teams of agents because teams need owners and of agents because teams need owners and of agents because teams need owners and status and receipts and review. And if status and receipts and review. And if status and receipts and review. And if you don't tackle that, something gets you don't tackle that, something gets you don't tackle that, something gets lost. Something drops through the lost. Something drops through the lost. Something drops through the cracks. And you have household cracks. And you have household cracks. And you have household consequences for that. You have work consequences for that. You have work consequences for that. You have work consequences for that. If we go back to consequences for that. If we go back to consequences for that. If we go back to earlier in this video, I talked about earlier in this video, I talked about earlier in this video, I talked about the idea that we are the hallway between the idea that we are the hallway between the idea that we are the hallway between AI agents. Open engine is what I built AI agents. Open engine is what I built AI agents. Open engine is what I built to make that hallway into software to make that hallway into software to make that hallway into software transparent and smooth and easy so we transparent and smooth and easy so we transparent and smooth and easy so we humans stop carrying all of that load. humans stop carrying all of that load. humans stop carrying all of that load. It it's not up to us to be the invisible It it's not up to us to be the invisible It it's not up to us to be the invisible laborers for our AIS. It is up to good laborers for our AIS. It is up to good laborers for our AIS. It is up to good frameworks and systems to let work move frameworks and systems to let work move frameworks and systems to let work move from where it started to where it needs from where it started to where it needs from where it started to where it needs to go. With enough context, the next to go. With enough context, the next to go. With enough context, the next loop can do something useful. The test loop can do something useful. The test loop can do something useful. The test for Open Engine is really simple. Can for Open Engine is really simple. Can for Open Engine is really simple. Can the work get out of your chat? Can it the work get out of your chat? Can it the work get out of your chat? Can it get out of your chat? Can it carry with get out of your chat? Can it carry with get out of your chat? Can it carry with it the sources that it has to do the it the sources that it has to do the it the sources that it has to do the work? Can it respect limits? And can it work? Can it respect limits? And can it work? Can it respect limits? And can it come back and say this is what I did and come back and say this is what I did and come back and say this is what I did and this is what I didn't do and this is the this is what I didn't do and this is the this is what I didn't do and this is the receipt because if the answer is yes receipt because if the answer is yes receipt because if the answer is yes with open engine even in a small way with open engine even in a small way with open engine even in a small way then the entire agent conversation then the entire agent conversation then the entire agent conversation changes for you. You stop asking only changes for you. You stop asking only changes for you. You stop asking only what one AI can do in a session. You what one AI can do in a session. You what one AI can do in a session. You start asking what kind of work your start asking what kind of work your start asking what kind of work your whole system can carry without you being whole system can carry without you being whole system can carry without you being the messenger. That's the promise of

the messenger. That's the promise of the messenger. That's the promise of open engine. Not agents replacing our open engine. Not agents replacing our open engine. Not agents replacing our judgment, not one agent to rule them judgment, not one agent to rule them judgment, not one agent to rule them all, but all of the mess of our AI all, but all of the mess of our AI all, but all of the mess of our AI systems just stitched together with a systems just stitched together with a systems just stitched together with a clean framework so agents can carry work clean framework so agents can carry work clean framework so agents can carry work to the point where judgment is needed to the point where judgment is needed to the point where judgment is needed and not bother us for the annoying and not bother us for the annoying and not bother us for the annoying handoffs along the way. That's the handoffs along the way. That's the handoffs along the way. That's the version of autonomy I actually want. version of autonomy I actually want. version of autonomy I actually want. That's what I've been living with. That's what I've been living with. That's what I've been living with. That's what I'm sharing with you today. That's what I'm sharing with you today. That's what I'm sharing with you today. If you want that, I have the full guide If you want that, I have the full guide If you want that, I have the full guide up. Uh the Substack community gets it up. Uh the Substack community gets it up. Uh the Substack community gets it first. We have an active Slack where first. We have an active Slack where first. We have an active Slack where people are building on this. I encourage people are building on this. I encourage people are building on this. I encourage you to join that as well. Uh and you can you to join that as well. Uh and you can you to join that as well. Uh and you can see exactly the the demo that I showed see exactly the the demo that I showed see exactly the the demo that I showed you and also a full guide that you can you and also a full guide that you can you and also a full guide that you can hand to your AI to get this started hand to your AI to get this started hand to your AI to get this started today. And remember, so that's open today. And remember, so that's open today. And remember, so that's open engine. If you have trouble with engine. If you have trouble with engine. If you have trouble with handoffs, as I have had and as so many handoffs, as I have had and as so many handoffs, as I have had and as so many folks I have had have struggled with, folks I have had have struggled with, folks I have had have struggled with, this is for you. Go make your life this is for you. Go make your life this is for you. Go make your life easier. Go get the headaches out of the easier. Go get the headaches out of the easier. Go get the headaches out of the way. Go stop copying and pasting and way. Go stop copying and pasting and way. Go stop copying and pasting and just make make that all go away. Get just make make that all go away. Get just make make that all go away. Get Open engine and get it sorted out so Open engine and get it sorted out so Open engine and get it sorted out so that you save hours. That's actually my that you save hours. That's actually my that you save hours. That's actually my personal goal is I want less wasted personal goal is I want less wasted personal goal is I want less wasted human hours. I want us to spend less human hours. I want us to spend less human hours. I want us to spend less hours on needless agent coordination.

hours on needless agent coordination. hours on needless agent coordination. That would be my success. So if you see That would be my success. So if you see That would be my success. So if you see this and you're like, "Oh my gosh, this this and you're like, "Oh my gosh, this this and you're like, "Oh my gosh, this is for me." Tell me in the chat in the is for me." Tell me in the chat in the is for me." Tell me in the chat in the comments what you are going to use this comments what you are going to use this comments what you are going to use this for. Tell me which AI systems you're for. Tell me which AI systems you're for. Tell me which AI systems you're going to coordinate. Tell me if you're going to coordinate. Tell me if you're going to coordinate. Tell me if you're working with agents, with a team or with working with agents, with a team or with working with agents, with a team or with your partner. Uh, whatever it is. And your partner. Uh, whatever it is. And your partner. Uh, whatever it is. And and you want a framework that makes the and you want a framework that makes the and you want a framework that makes the AI and the human actually work together AI and the human actually work together AI and the human actually work together and you're going to use this to solve and you're going to use this to solve and you're going to use this to solve it. Tell me what it is. One, we'll it. Tell me what it is. One, we'll it. Tell me what it is. One, we'll improve Open Engine and continue to make improve Open Engine and continue to make improve Open Engine and continue to make it better as you give us feedback. But it better as you give us feedback. But it better as you give us feedback. But but two, I want to know how much pain but two, I want to know how much pain but two, I want to know how much pain we're going after here because that's we're going after here because that's we're going after here because that's what I'm interested in. I'm interested what I'm interested in. I'm interested what I'm interested in. I'm interested in AI that kills our pain. And this is a in AI that kills our pain. And this is a in AI that kills our pain. And this is a pain that has arisen since AI agents got pain that has arisen since AI agents got pain that has arisen since AI agents got more capable and it's just gotten worse. more capable and it's just gotten worse. more capable and it's just gotten worse. And so in the era of openclaw, in the And so in the era of openclaw, in the And so in the era of openclaw, in the era of codecs, in the era of claude, we era of codecs, in the era of claude, we era of codecs, in the era of claude, we need something like open engine. I hope need something like open engine. I hope need something like open engine. I hope you've had fun seeing this demo.