- Anthropic has rebranded its developer offering as the Claude Developer Platform, evolving from a simple API to a comprehensive suite of tools and SDKs that enable building sophisticated AI agents.
- The core philosophy for building agents is to "unhobble the model" by providing powerful tools and less scaffolding, allowing the model more autonomy to choose actions and integrate new capabilities.
- The Claude Code SDK serves as a general-purpose agent harness, simplifying the development of autonomous agentic loops for various tasks, not just coding, and is recommended for rapid prototyping.
Building the future of agents with Claude
- The
Claude Developer Platformencompasses Anthropic's APIs, SDKs, documentation, and console experiences, serving as the complete ecosystem for building on Claude. - An
agentis defined by Anthropic as a system where the model autonomously chooses and calls tools, handles results, and decides on the next steps in a dynamic workflow. - As model intelligence increases, agents require less
scaffoldingand predefined workflows; excessive guardrails can actually limit the model's ability to leverage new intelligence. - The
Claude Code SDK, despite its name, is designed as a general-purposeagent harnessthat automates the agentic loop and tool calling, providing an out-of-the-box solution for prototyping agents. - Key tools available include
server-side web searchandweb fetch, which enable models to perform deep research tasks autonomously by calling and processing external information. - New context management features allow models to intelligently remove older, unnecessary
tool callsfrom thecontext windowto improve focus, whiletombstoningensures basic awareness of removed items. - An
agentic memoryfeature allows the model to take and review notes during longer tasks, enabling it to learn and improve performance over multiple runs. - Future developments aim for higher-order
abstractionspaired withobservabilitytools to enable self-improving and continuously enhancing agent outcomes, along with giving Claude a "computer" for persistent state and file management. - When embarking on agent projects, developers should clearly articulate the business value (e.g., saving engineering hours, reducing manual work) to effectively define the scope.
Anthropic API — The original, simpler interface for accessing Anthropic's models, now evolved into the Claude Developer Platform.
Claude Developer Platform — Anthropic's comprehensive offering for developers, including APIs, SDKs, documentation, and developer console experiences.
agent — An AI system where the language model autonomously selects and uses tools, processes outcomes, and determines subsequent actions.
agentic pattern — A design approach for AI applications that emphasizes the model's autonomy in decision-making and tool utilization.
scaffolding — Predefined structures or guardrails that guide an AI model's behavior, which can become restrictive as models become more capable.
agent harness — A framework or toolkit that wraps an AI model to manage the iterative agentic loop of tool calling, response handling, and decision making.
context window — The maximum amount of text (tokens) a language model can process or "remember" at any given time.
tool call — An instruction from a language model to use an external function or API, often followed by processing the returned results.
tombstone — A placeholder or brief note left in the model's context when a larger piece of information (like old tool results) is removed to conserve tokens, indicating what was there.
observability — The ability to understand the internal state of a system (like an agent's long-running task) by examining its outputs and logs, crucial for auditing and tuning.
Because as a developer, like my creativity ends at some point, I can only think of so many use cases, but the model, like anything, anything somebody comes with, the model will figure out a way to do that thing. Hey, I'm Alex, I lead Claude Relations here at Anthropic. Today, we're talking about building the future of agents with Claude, and I'm joined by my colleagues. I'm Brad. I run the PM team on the Claude Developer Platform. I'm Caitlin. I lead the engineering team for the Claude Developer Platform. Let's talk about the Claude Developer Platform. Yeah, start with that. Let's start with that. You speak called the Anthropic API. Yeah, we just went through a big name change. Yeah. Can you walk me through why we made that change and also what this new platform is and what it encompasses? Yeah, totally. So the Claude Developer Platform Reeling Compuses are APIs, RSDKs, our documentation, all of our experiences within the console, and really everything that a developer needs to actually build on top of Claude. We're really humbled, proud to serve some really awesome customers around the world who are trying to, what we like to say, raise the ceiling of intelligence using Claude, and the platform really enables them to do that. And I would say one of my favorite parts about it is the platform doesn't just serve customers externally. The platform actually serves our internal product. So we love telling people like Claude Code, for example, is actually built directly on our public platform. I see. Yeah, I mean, I think when we started, we were just the Anthropic API. It's very simple access to the model. But over the last year, so we've added so many features to it. We added prompt caching. We added a whole separate batch of API. We added web search, a web fetch. This context management support, the code execution. So all these tools, now this kind of, we feel like, yeah, it's aspirationally, it's a platform now. I see. So there's just a lot more to it now. It's evolved in a pretty drastic way over the past year. Yeah, yeah, better than anything. Yeah. And I think that's what developers were sort of calling it anyway. Yeah. So it's always natural to just sort of go with what developers were saying. Right. We're a little late to the game there. So it's OK. It's had it right. It's OK. It's OK. We've made our mends. One of the cool things you can do now, as we're moving from the sort of chat model to maybe this more authentic future, is building agents as part of this developer platform. Before we get into how we're actually doing that, on the platform, can we talk about what even is an agent to begin with? Yeah. I mean, agents is, it's almost sort of a buzzword, right? Yeah. Everybody you talk to now is building agents. And whenever industry tech term gets to that level, the definition gets very gray. Because everything everybody builds as an agent. But inthropic, what we really think about agent is where the model is taking some autonomy to be able to choose what tools to call, to call those tools, to handle the results, and choose the next step. So as a foundational research lab, leaning into the model and what it's reasoning, how it decides what to do, we think that's a really important element of what an agent is. So it's kind of like the aspect of it being autonomous in some sense. Yeah. Yeah. Yeah. Charlie, I mean, I think there's also, I mean, we have customers doing really useful workflows where they're sort of pre-defining the path that Clod should walk, and that is a super useful thing to do. But what's nice about the agentic thing is as the model gets better, every couple of months, we release a new model. And with a true agentic pattern, those services are just going to get better. Where if you build a workflow with a lot of scaffolding in it, you kind of put bounds on the model, which is maybe OK in some use cases. But that means that you may not take advantage of the next level of intelligence that a next model release gets. Yeah. So it seems like there's this interesting trend with agents at least over the past six to 12 months, where like you've said, the scaffolding has been a bit of a hindrance, and maybe we're dropping some of that. Can you explain the intuitions behind that around, is this actually the future? Is like we give less and less things to the model? Yeah. I mean, I think over time, what we're seeing is the scaffolding the model needs to be able to accomplish tasks. It's needing less as the level of intelligence of the model goes up. And we believe it's going to keep going up that basically the model has more contextual understanding of the high level task that it's trying to accomplish. So therefore, it doesn't need as many sort of guardrails. And in fact, those guardrails, in some cases, become like a liability to have. We've had customers try out new models and say, oh, well, it's actually only just a little bit better. And then we kind of look into it with them about what's going on. And it turns out, well, yeah, they were constraining it in ways that makes it harder for them to see the intelligence of the model. Does this match what we see in the field with our customers where they're also following these same trends? I know at the limit, we have customers exploring all sorts of innovative techniques for managing Claude. Yeah, totally. And there's actually a lot of discourse about this right now, right? What is an agent and what does it need? What do you need to build? And there are people saying, it's just a while loop. We don't have to try that hard. And I think ultimately, there's been a lot of evolution of frameworks that people are putting around the model that are helping them orchestrate their agents, try to get the most out of the model. And I think what the industry is maybe kind of circling around is a lot of that has become maybe too heavy and maybe too opinionated. Which is why you kind of get the people coming back to you like, it's just a while loop and that is all you need. And I think what we're trying to do there is to say maybe in a lot of ways it is a while loop but the things we can more uniquely do to help people get the most out of the model is a lot of those tools, those features, and otherwise. And so what we want to do is put frameworks and tools and platform out there that is opinionated to some extent on how people should use those tools. But it's not this super heavy framework that really like to Brad's point gets in the way of what the model is ultimately trying to do. So it's strike the right balance. It's like, we've seen what a lot of people tried to do so we know we can be opinionated there but we want to be lightweight in the way that we're doing that and make sure that the real thing we're doing is helping you get the most out of the model without bogging you down in some super heavy framework. Right, so would you describe part of the strategy here then as providing these auxiliary tools and things that we can give to the model but we're not necessarily placing the bumper on the model itself or something. Yeah, we think about it as like, how do you unhobble the model? The model already has a lot of capabilities. And in fact, I'm convinced that even if you take your current generation of models, there's way more intelligence in there than we've been able to unlock. But anyway, that intuition is if you just give the model like the tools it needs and set it free, let it be able to use those in the right way, you'll get great results. And I think like a good example of that is we launched this server side web search tool and web fetch tools. And it's been interesting to watch customers use those and all we did real, I mean, it's a very minimal prompt that we have, we just give it the web search tool. And like all of a sudden deep research tasks are almost completely done with just turning on that switch on the API because the model will call that tool, it'll look at its results, it'll say consider it and say, okay, maybe I need to call, do these other searches and then all that fourth link you return, that's the great one. It'll do a web fetch on that link and bring that data back. And really all that very autonomously on its own kind of deciding. Right, I think it's almost kind of like an interesting shift in like where the intelligence of a system is being applied. From like the developer having to apply their intelligence to guiding towards like the model now, to get out. And it's so exciting with the model does it because as a developer like my creativity ends at some point I can only think of so many use cases. But the model like anything, anything somebody comes with, the model will figure out a way to do that thing. So it's great, great to unhobble the model. Yeah, so if I'm a developer today and I'm getting started building with the developer platform, what do you recommend? Where are some best practices or ways for me to get started? Yeah, so super tactically actually the number one thing that we recommend right now is the Claude Code SDK. And what's really really interesting about the Claude Code SDK is we essentially built an agent harness, an agentech harness around the model to run that loop, right, and automate a lot of that tool calling and otherwise feature use. And obviously originally was built for coding purposes. And what the team really quickly figured out was actually this is like an excellent general purpose agentech harness. And so what the SDK does is it gives people a perfect out of the box solution to actually just start prototyping agents without having to go and build the loop with all the tool calling and otherwise. It's built on top of the messages API and all those same tools that we're mentioning. But it kind of gives you that really great starting place right out of the box. Right, I feel like this is a pretty common misconception at least when I talk to developers about the Claude Code SDK. So I'm not building a coding application. Why would I want to use this? But you can kind of remove the coding specific parts. I mean, I think that's a great example of what we were talking about removing scaffolding on the model. It's like once we got done removing things from Claude Code to really unhobble the model, it turns out there was nothing coding left. When you remove everything else, then it's just an agentic loop. And you're really a minimalistic thing to give Claude access to a file system, to a set of Linux command line tools, to the ability to write code and execute that code. So those are all very generic capabilities that turns out could solve a wide variety of problems. Right, yeah, I feel like something I've been running up to and my own side projects and also seeing with projects within Anthropic is before the Claude Code SDK, everybody's implementing some form of managing prompt caching or some form of managing their tool calls and that loop. Right, right. And now it's like, oh, just started this point. I then build from there. You start a little bit higher up. Yeah, yeah, yeah. So it's like a further level abstraction. I think that's super interesting. I think the other really interesting thing to think about, especially for businesses looking at agencies, like what use case to go target. So thinking beyond the technology, like what is the actual problem to go solve? And I think we see a lot of customers and doing a lot of things, we love all of it. But where the biggest impacts are is where the customer has thought hard about what's the business value of this. Will it actually save this many engineering hours or will it help us remove this much manual work or whatnot? And being able to articulate what you expect the outcome of the agent project to be, I think is really helpful in defining the scope of the agent. Right. And time back one more time to the SDK. So it seems like it's been really, really useful for like individual developers, like myself, you know, starting out and just wanting to get hacking on something really fast for these customers, for enterprises that are actually trying to get real business value of these things. Should they be using the SDK? Is it ready for them? Is it ready for skilled use like that? Yeah. So I think in a lot of ways it is. In a lot of ways, if you are in a spot where you can deploy that runtime, essentially that's what you get out of the SDK is an agentic loop runtime. You can go and deploy that runtime wherever you want whenever you're ready to do so. But I think what we're really trying to do is take the spirit of what the SDK unlocks for people. Like go kind of up to that higher order abstraction where we give you the loop, we give you a lot of the tool calling in an automated way and say, how can we learn from that and give people out of the box solutions that like at scale will really be able to solve for their use cases. And I think that's a lot of where we're kind of trying to go with our roadmap throughout the rest of the year. And one really important bit when we think about that is if the entire goal here is to help our users like really raise that ceiling of intelligence, get the absolute best outcome out of the models, then higher order abstractions are not just make it easier because you don't have to write all that code yourself. It's actually like how can we like really truly help you get the best outcome because we were in the room with research, we're in the room with inference, like we know how to make sure that our abstractions, our agentic loop is going to be extremely powerful and extremely good at working with Claude. And the last thing that I would add in there is especially as these things get longer running and as we provide more and more tooling to help people get at those longer running tasks, another big problem that our users, we know we're going to keep trying to solve is observability within those longer running tasks. And so that's one of the most common things that comes up for folks is I have these long running tasks I'm trying to get these really great outcomes but I might need to do some steering where I might need to tune my prompt or I might need to think about tool calling a little differently and that's something that we know we can give people that observability through the platform over time and that's another big area of focus for us. Okay, that's really interesting. I mean, this has been a huge issue that's starting to come to a head with agents. I think so especially as you trust them to go work in some other applications in the background. How do you make sure they're actually doing the right thing and then if you're deploying them. Yeah, how do you audit it? Like if we're gonna give some level of autonomy to the system, there needs to be a way to audit it and like make sure the right things are happening so that you can tune things and whatnot. So I think observability is really a key piece of this. And putting a pin there, I wanna ask a question on just like the future of how we're gonna address that. Before I do, is there other tools that exist right now that folks should be aware of when they're getting started with the developer platform? Things you've found helpful or useful? Yeah, I mean, I think there's a, so we mentioned web search and web fetch. I think another big thing that we're seeing is customers have to do, right now have to do a lot of work to manage the context window. So by default, Claude has 200K tokens of context. We have a million token available now in beta on SANA, which is great, but even a million there's a limit there. And what many customers have told us is that they get better outputs, higher intelligence if they even use a smaller part of the context. And so we've done, we have a couple of cool features that are just coming out to help developers manage that context. So in these agentec loops, a lot of times you're doing 10, 15, 100 tool calls, edit this file or look up data in this database or send this email. And each of those tool calls takes up 100, 200, 1000 tokens. And so we have this cool feature that lets you, lets the model actually remove some of the older tool calls that are not needed anymore. Interesting. And that gives just like you, if you declutter your desk and declutter your notebook, like you can focus a little bit better. So if you declutter the prompt, actually, the model can actually focus a little bit better. Interesting. So, okay, we're moving unnecessary context. Is there a risk that we remove unnecessary context? Yeah, yeah, yeah. So we have some guard rails and some bounds around it. So you don't, but the general rule is if you remove, we try to remove the tools that are like several turns back. Okay. The model's already made decisions based on those tools. But if you, yeah, I was playing with it recently and I removed the tools that it was just called. And it's, oh, my tool results are gone. I don't know what to do. And then the, but the model that's on it doesn't give up. Like it's like, I'm just gonna call this tool again, you know? Yeah, yeah, yeah. But yeah, so generally we have put some bounds on that because of that experience. So we do preserve the most recent set of tools. I see. And then the other cool thing we do is, tombstone it. So by that we mean, when we remove the tool calls, we put a note in there to the model. I say, oh, the tool results for the search call are what we're here. Oh, okay. We've been removed. The model's not completely like memory wipes. Exactly. I think we found the model does better if we just give it a little more context about what is happening. Right. And so that's a key feature. Right. And the other one is this like a kind of agentic memory feature that we've added. And there we have seen that the model does, like right now if you give a task to the model, say a deep research task or play Pokemon or whatnot, like the model does about the same every time it runs. But if you give a human a task, like the fifth time the human does a task, like they do it like way better. Because they've learned, okay, if I'm gonna do this search, okay, probably the Wikipedia site is better than this other site or whatever. Like they learn which thing so they get better over time. So we've given this memory tool to the model now so that the model can actually take some notes while it's going and say, oh, I realize that this website maybe isn't the right one or if I'm doing a search, it should be like this or if I'm looking up, I should use this database, not that database or whatnot. And it makes those notes. And then when it's stumped, it can actually go back and review its notes and say, okay, like, oh, I'm starting this task, let me go read the notes so I can figure it out. Cool. So we're handling all of that for the developer. Yeah, yeah, well, we're giving the model like this core capability to do memory. And right now we're letting the developer manage the memory because different developers like they might want to store it in some cloud storage or they might want to store it somewhere else. So we're letting developers figure out exactly where to store the memory that way they have more control over that. But exposing the tool. But exposing the tool, I would say. So going back again to a roadmap question here. So it sounds like there's a ton of new features that we've recently launched. There's a lot of momentum and now there's other offerings as well like the Claude Code SDK and things coming out soon. What are you most excited about, Caitlin? What's the future looking like here in the next six, 12 months? Yeah. So we talked a little bit about these higher orders of extraction where we can really just make it as simple as possible for you to get the absolute best outcomes out of Claude. And we want to pair that with the observability that we talked about so that you can really like, see the data and take those insights from those longer running tasks. And if you combine these things together and start to think about some of the capabilities like memory that Brad just talked about, you can really start to see this flywheel where over time we're not just able to help you get the best outcomes out of Claude, but we can help you get self-improving and continuously improving outcomes out of Claude. And that to me is kind of the galaxy brain magic of the roadmap is get to a point where we have people coming to us, they're building on Claude, they have their tasks, they know what they're trying to do. And they get these really like aha moments where over time, it's getting better and better and better. And that's kind of the biggest thing that in everything that we're doing, we're trying to make sure we're going after. That's awesome. Yeah, I guess I'd have to say, I'm always excited about model launches. It's like Christmas, like what will be possible now? So I love playing with the model launches that they come out, just unlocks more use cases. Some use cases that we've been working hard on and trying to improve, which is satisfying to see. But also some things, I wouldn't know idea of the model would be able to do this thing. Now, it's also asking pictures so much better or whatever the things. You can be very important things. But beyond that, the other thing I'm really excited about is we're in the early stages of giving Claude a computer. I think about if we hire an employee here at Anthropic and we welcome him here's your first day, but we don't give them a computer, they would not be very successful at Anthropic. So right now, essentially everybody is using Claude and it doesn't have a computer. So I'm really excited about giving Claude a computer. And you see the very baby steps of that with the code execution tool, where the model can write code executed on the VM and get the results back. So it can zoom in on images or take a Excel spreadsheet and create amazing data analysis with charts and graphs. And that's just the baby step. What if I had a persistent computer that was always there and it could organize the files in there the way it needed and get the tools set up the way it wanted. And I just think there's a lot of headroom to that scenario. Yeah, and I guess that all ties back into this unhobbling. It gets to it. Exactly. All about the modeling model. That's exactly. Just give, give Claude the tools. Yeah. Well, I'm excited for that future. Thanks so much for this conversation. All right, cool. Thank you.
TL;DR
- Anthropic has rebranded its developer offering as the Claude Developer Platform, evolving from a simple API to a comprehensive suite of tools and SDKs that enable building sophisticated AI agents.
- The core philosophy for building agents is to "unhobble the model" by providing powerful tools and less scaffolding, allowing the model more autonomy to choose actions and integrate new capabilities.
- The Claude Code SDK serves as a general-purpose agent harness, simplifying the development of autonomous agentic loops for various tasks, not just coding, and is recommended for rapid prototyping.
Takeaways
- The
Claude Developer Platformencompasses Anthropic's APIs, SDKs, documentation, and console experiences, serving as the complete ecosystem for building on Claude. - An
agentis defined by Anthropic as a system where the model autonomously chooses and calls tools, handles results, and decides on the next steps in a dynamic workflow. - As model intelligence increases, agents require less
scaffoldingand predefined workflows; excessive guardrails can actually limit the model's ability to leverage new intelligence. - The
Claude Code SDK, despite its name, is designed as a general-purposeagent harnessthat automates the agentic loop and tool calling, providing an out-of-the-box solution for prototyping agents. - Key tools available include
server-side web searchandweb fetch, which enable models to perform deep research tasks autonomously by calling and processing external information. - New context management features allow models to intelligently remove older, unnecessary
tool callsfrom thecontext windowto improve focus, whiletombstoningensures basic awareness of removed items. - An
agentic memoryfeature allows the model to take and review notes during longer tasks, enabling it to learn and improve performance over multiple runs. - Future developments aim for higher-order
abstractionspaired withobservabilitytools to enable self-improving and continuously enhancing agent outcomes, along with giving Claude a "computer" for persistent state and file management. - When embarking on agent projects, developers should clearly articulate the business value (e.g., saving engineering hours, reducing manual work) to effectively define the scope.
Vocabulary
Anthropic API — The original, simpler interface for accessing Anthropic's models, now evolved into the Claude Developer Platform.
Claude Developer Platform — Anthropic's comprehensive offering for developers, including APIs, SDKs, documentation, and developer console experiences.
agent — An AI system where the language model autonomously selects and uses tools, processes outcomes, and determines subsequent actions.
agentic pattern — A design approach for AI applications that emphasizes the model's autonomy in decision-making and tool utilization.
scaffolding — Predefined structures or guardrails that guide an AI model's behavior, which can become restrictive as models become more capable.
agent harness — A framework or toolkit that wraps an AI model to manage the iterative agentic loop of tool calling, response handling, and decision making.
context window — The maximum amount of text (tokens) a language model can process or "remember" at any given time.
tool call — An instruction from a language model to use an external function or API, often followed by processing the returned results.
tombstone — A placeholder or brief note left in the model's context when a larger piece of information (like old tool results) is removed to conserve tokens, indicating what was there.
observability — The ability to understand the internal state of a system (like an agent's long-running task) by examining its outputs and logs, crucial for auditing and tuning.
Transcript
Because as a developer, like my creativity ends at some point, I can only think of so many use cases, but the model, like anything, anything somebody comes with, the model will figure out a way to do that thing. Hey, I'm Alex, I lead Claude Relations here at Anthropic. Today, we're talking about building the future of agents with Claude, and I'm joined by my colleagues. I'm Brad. I run the PM team on the Claude Developer Platform. I'm Caitlin. I lead the engineering team for the Claude Developer Platform. Let's talk about the Claude Developer Platform. Yeah, start with that. Let's start with that. You speak called the Anthropic API. Yeah, we just went through a big name change. Yeah. Can you walk me through why we made that change and also what this new platform is and what it encompasses? Yeah, totally. So the Claude Developer Platform Reeling Compuses are APIs, RSDKs, our documentation, all of our experiences within the console, and really everything that a developer needs to actually build on top of Claude. We're really humbled, proud to serve some really awesome customers around the world who are trying to, what we like to say, raise the ceiling of intelligence using Claude, and the platform really enables them to do that. And I would say one of my favorite parts about it is the platform doesn't just serve customers externally. The platform actually serves our internal product. So we love telling people like Claude Code, for example, is actually built directly on our public platform. I see. Yeah, I mean, I think when we started, we were just the Anthropic API. It's very simple access to the model. But over the last year, so we've added so many features to it. We added prompt caching. We added a whole separate batch of API. We added web search, a web fetch. This context management support, the code execution. So all these tools, now this kind of, we feel like, yeah, it's aspirationally, it's a platform now. I see. So there's just a lot more to it now. It's evolved in a pretty drastic way over the past year. Yeah, yeah, better than anything. Yeah. And I think that's what developers were sort of calling it anyway. Yeah. So it's always natural to just sort of go with what developers were saying. Right. We're a little late to the game there. So it's OK. It's had it right. It's OK. It's OK. We've made our mends. One of the cool things you can do now, as we're moving from the sort of chat model to maybe this more authentic future, is building agents as part of this developer platform. Before we get into how we're actually doing that, on the platform, can we talk about what even is an agent to begin with? Yeah. I mean, agents is, it's almost sort of a buzzword, right? Yeah. Everybody you talk to now is building agents. And whenever industry tech term gets to that level, the definition gets very gray. Because everything everybody builds as an agent. But inthropic, what we really think about agent is where the model is taking some autonomy to be able to choose what tools to call, to call those tools, to handle the results, and choose the next step. So as a foundational research lab, leaning into the model and what it's reasoning, how it decides what to do, we think that's a really important element of what an agent is. So it's kind of like the aspect of it being autonomous in some sense. Yeah. Yeah. Yeah. Charlie, I mean, I think there's also, I mean, we have customers doing really useful workflows where they're sort of pre-defining the path that Clod should walk, and that is a super useful thing to do. But what's nice about the agentic thing is as the model gets better, every couple of months, we release a new model. And with a true agentic pattern, those services are just going to get better. Where if you build a workflow with a lot of scaffolding in it, you kind of put bounds on the model, which is maybe OK in some use cases. But that means that you may not take advantage of the next level of intelligence that a next model release gets. Yeah. So it seems like there's this interesting trend with agents at least over the past six to 12 months, where like you've said, the scaffolding has been a bit of a hindrance, and maybe we're dropping some of that. Can you explain the intuitions behind that around, is this actually the future? Is like we give less and less things to the model? Yeah. I mean, I think over time, what we're seeing is the scaffolding the model needs to be able to accomplish tasks. It's needing less as the level of intelligence of the model goes up. And we believe it's going to keep going up that basically the model has more contextual understanding of the high level task that it's trying to accomplish. So therefore, it doesn't need as many sort of guardrails. And in fact, those guardrails, in some cases, become like a liability to have. We've had customers try out new models and say, oh, well, it's actually only just a little bit better. And then we kind of look into it with them about what's going on. And it turns out, well, yeah, they were constraining it in ways that makes it harder for them to see the intelligence of the model. Does this match what we see in the field with our customers where they're also following these same trends? I know at the limit, we have customers exploring all sorts of innovative techniques for managing Claude. Yeah, totally. And there's actually a lot of discourse about this right now, right? What is an agent and what does it need? What do you need to build? And there are people saying, it's just a while loop. We don't have to try that hard. And I think ultimately, there's been a lot of evolution of frameworks that people are putting around the model that are helping them orchestrate their agents, try to get the most out of the model. And I think what the industry is maybe kind of circling around is a lot of that has become maybe too heavy and maybe too opinionated. Which is why you kind of get the people coming back to you like, it's just a while loop and that is all you need. And I think what we're trying to do there is to say maybe in a lot of ways it is a while loop but the things we can more uniquely do to help people get the most out of the model is a lot of those tools, those features, and otherwise. And so what we want to do is put frameworks and tools and platform out there that is opinionated to some extent on how people should use those tools. But it's not this super heavy framework that really like to Brad's point gets in the way of what the model is ultimately trying to do. So it's strike the right balance. It's like, we've seen what a lot of people tried to do so we know we can be opinionated there but we want to be lightweight in the way that we're doing that and make sure that the real thing we're doing is helping you get the most out of the model without bogging you down in some super heavy framework. Right, so would you describe part of the strategy here then as providing these auxiliary tools and things that we can give to the model but we're not necessarily placing the bumper on the model itself or something. Yeah, we think about it as like, how do you unhobble the model? The model already has a lot of capabilities. And in fact, I'm convinced that even if you take your current generation of models, there's way more intelligence in there than we've been able to unlock. But anyway, that intuition is if you just give the model like the tools it needs and set it free, let it be able to use those in the right way, you'll get great results. And I think like a good example of that is we launched this server side web search tool and web fetch tools. And it's been interesting to watch customers use those and all we did real, I mean, it's a very minimal prompt that we have, we just give it the web search tool. And like all of a sudden deep research tasks are almost completely done with just turning on that switch on the API because the model will call that tool, it'll look at its results, it'll say consider it and say, okay, maybe I need to call, do these other searches and then all that fourth link you return, that's the great one. It'll do a web fetch on that link and bring that data back. And really all that very autonomously on its own kind of deciding. Right, I think it's almost kind of like an interesting shift in like where the intelligence of a system is being applied. From like the developer having to apply their intelligence to guiding towards like the model now, to get out. And it's so exciting with the model does it because as a developer like my creativity ends at some point I can only think of so many use cases. But the model like anything, anything somebody comes with, the model will figure out a way to do that thing. So it's great, great to unhobble the model. Yeah, so if I'm a developer today and I'm getting started building with the developer platform, what do you recommend? Where are some best practices or ways for me to get started? Yeah, so super tactically actually the number one thing that we recommend right now is the Claude Code SDK. And what's really really interesting about the Claude Code SDK is we essentially built an agent harness, an agentech harness around the model to run that loop, right, and automate a lot of that tool calling and otherwise feature use. And obviously originally was built for coding purposes. And what the team really quickly figured out was actually this is like an excellent general purpose agentech harness. And so what the SDK does is it gives people a perfect out of the box solution to actually just start prototyping agents without having to go and build the loop with all the tool calling and otherwise. It's built on top of the messages API and all those same tools that we're mentioning. But it kind of gives you that really great starting place right out of the box. Right, I feel like this is a pretty common misconception at least when I talk to developers about the Claude Code SDK. So I'm not building a coding application. Why would I want to use this? But you can kind of remove the coding specific parts. I mean, I think that's a great example of what we were talking about removing scaffolding on the model. It's like once we got done removing things from Claude Code to really unhobble the model, it turns out there was nothing coding left. When you remove everything else, then it's just an agentic loop. And you're really a minimalistic thing to give Claude access to a file system, to a set of Linux command line tools, to the ability to write code and execute that code. So those are all very generic capabilities that turns out could solve a wide variety of problems. Right, yeah, I feel like something I've been running up to and my own side projects and also seeing with projects within Anthropic is before the Claude Code SDK, everybody's implementing some form of managing prompt caching or some form of managing their tool calls and that loop. Right, right. And now it's like, oh, just started this point. I then build from there. You start a little bit higher up. Yeah, yeah, yeah. So it's like a further level abstraction. I think that's super interesting. I think the other really interesting thing to think about, especially for businesses looking at agencies, like what use case to go target. So thinking beyond the technology, like what is the actual problem to go solve? And I think we see a lot of customers and doing a lot of things, we love all of it. But where the biggest impacts are is where the customer has thought hard about what's the business value of this. Will it actually save this many engineering hours or will it help us remove this much manual work or whatnot? And being able to articulate what you expect the outcome of the agent project to be, I think is really helpful in defining the scope of the agent. Right. And time back one more time to the SDK. So it seems like it's been really, really useful for like individual developers, like myself, you know, starting out and just wanting to get hacking on something really fast for these customers, for enterprises that are actually trying to get real business value of these things. Should they be using the SDK? Is it ready for them? Is it ready for skilled use like that? Yeah. So I think in a lot of ways it is. In a lot of ways, if you are in a spot where you can deploy that runtime, essentially that's what you get out of the SDK is an agentic loop runtime. You can go and deploy that runtime wherever you want whenever you're ready to do so. But I think what we're really trying to do is take the spirit of what the SDK unlocks for people. Like go kind of up to that higher order abstraction where we give you the loop, we give you a lot of the tool calling in an automated way and say, how can we learn from that and give people out of the box solutions that like at scale will really be able to solve for their use cases. And I think that's a lot of where we're kind of trying to go with our roadmap throughout the rest of the year. And one really important bit when we think about that is if the entire goal here is to help our users like really raise that ceiling of intelligence, get the absolute best outcome out of the models, then higher order abstractions are not just make it easier because you don't have to write all that code yourself. It's actually like how can we like really truly help you get the best outcome because we were in the room with research, we're in the room with inference, like we know how to make sure that our abstractions, our agentic loop is going to be extremely powerful and extremely good at working with Claude. And the last thing that I would add in there is especially as these things get longer running and as we provide more and more tooling to help people get at those longer running tasks, another big problem that our users, we know we're going to keep trying to solve is observability within those longer running tasks. And so that's one of the most common things that comes up for folks is I have these long running tasks I'm trying to get these really great outcomes but I might need to do some steering where I might need to tune my prompt or I might need to think about tool calling a little differently and that's something that we know we can give people that observability through the platform over time and that's another big area of focus for us. Okay, that's really interesting. I mean, this has been a huge issue that's starting to come to a head with agents. I think so especially as you trust them to go work in some other applications in the background. How do you make sure they're actually doing the right thing and then if you're deploying them. Yeah, how do you audit it? Like if we're gonna give some level of autonomy to the system, there needs to be a way to audit it and like make sure the right things are happening so that you can tune things and whatnot. So I think observability is really a key piece of this. And putting a pin there, I wanna ask a question on just like the future of how we're gonna address that. Before I do, is there other tools that exist right now that folks should be aware of when they're getting started with the developer platform? Things you've found helpful or useful? Yeah, I mean, I think there's a, so we mentioned web search and web fetch. I think another big thing that we're seeing is customers have to do, right now have to do a lot of work to manage the context window. So by default, Claude has 200K tokens of context. We have a million token available now in beta on SANA, which is great, but even a million there's a limit there. And what many customers have told us is that they get better outputs, higher intelligence if they even use a smaller part of the context. And so we've done, we have a couple of cool features that are just coming out to help developers manage that context. So in these agentec loops, a lot of times you're doing 10, 15, 100 tool calls, edit this file or look up data in this database or send this email. And each of those tool calls takes up 100, 200, 1000 tokens. And so we have this cool feature that lets you, lets the model actually remove some of the older tool calls that are not needed anymore. Interesting. And that gives just like you, if you declutter your desk and declutter your notebook, like you can focus a little bit better. So if you declutter the prompt, actually, the model can actually focus a little bit better. Interesting. So, okay, we're moving unnecessary context. Is there a risk that we remove unnecessary context? Yeah, yeah, yeah. So we have some guard rails and some bounds around it. So you don't, but the general rule is if you remove, we try to remove the tools that are like several turns back. Okay. The model's already made decisions based on those tools. But if you, yeah, I was playing with it recently and I removed the tools that it was just called. And it's, oh, my tool results are gone. I don't know what to do. And then the, but the model that's on it doesn't give up. Like it's like, I'm just gonna call this tool again, you know? Yeah, yeah, yeah. But yeah, so generally we have put some bounds on that because of that experience. So we do preserve the most recent set of tools. I see. And then the other cool thing we do is, tombstone it. So by that we mean, when we remove the tool calls, we put a note in there to the model. I say, oh, the tool results for the search call are what we're here. Oh, okay. We've been removed. The model's not completely like memory wipes. Exactly. I think we found the model does better if we just give it a little more context about what is happening. Right. And so that's a key feature. Right. And the other one is this like a kind of agentic memory feature that we've added. And there we have seen that the model does, like right now if you give a task to the model, say a deep research task or play Pokemon or whatnot, like the model does about the same every time it runs. But if you give a human a task, like the fifth time the human does a task, like they do it like way better. Because they've learned, okay, if I'm gonna do this search, okay, probably the Wikipedia site is better than this other site or whatever. Like they learn which thing so they get better over time. So we've given this memory tool to the model now so that the model can actually take some notes while it's going and say, oh, I realize that this website maybe isn't the right one or if I'm doing a search, it should be like this or if I'm looking up, I should use this database, not that database or whatnot. And it makes those notes. And then when it's stumped, it can actually go back and review its notes and say, okay, like, oh, I'm starting this task, let me go read the notes so I can figure it out. Cool. So we're handling all of that for the developer. Yeah, yeah, well, we're giving the model like this core capability to do memory. And right now we're letting the developer manage the memory because different developers like they might want to store it in some cloud storage or they might want to store it somewhere else. So we're letting developers figure out exactly where to store the memory that way they have more control over that. But exposing the tool. But exposing the tool, I would say. So going back again to a roadmap question here. So it sounds like there's a ton of new features that we've recently launched. There's a lot of momentum and now there's other offerings as well like the Claude Code SDK and things coming out soon. What are you most excited about, Caitlin? What's the future looking like here in the next six, 12 months? Yeah. So we talked a little bit about these higher orders of extraction where we can really just make it as simple as possible for you to get the absolute best outcomes out of Claude. And we want to pair that with the observability that we talked about so that you can really like, see the data and take those insights from those longer running tasks. And if you combine these things together and start to think about some of the capabilities like memory that Brad just talked about, you can really start to see this flywheel where over time we're not just able to help you get the best outcomes out of Claude, but we can help you get self-improving and continuously improving outcomes out of Claude. And that to me is kind of the galaxy brain magic of the roadmap is get to a point where we have people coming to us, they're building on Claude, they have their tasks, they know what they're trying to do. And they get these really like aha moments where over time, it's getting better and better and better. And that's kind of the biggest thing that in everything that we're doing, we're trying to make sure we're going after. That's awesome. Yeah, I guess I'd have to say, I'm always excited about model launches. It's like Christmas, like what will be possible now? So I love playing with the model launches that they come out, just unlocks more use cases. Some use cases that we've been working hard on and trying to improve, which is satisfying to see. But also some things, I wouldn't know idea of the model would be able to do this thing. Now, it's also asking pictures so much better or whatever the things. You can be very important things. But beyond that, the other thing I'm really excited about is we're in the early stages of giving Claude a computer. I think about if we hire an employee here at Anthropic and we welcome him here's your first day, but we don't give them a computer, they would not be very successful at Anthropic. So right now, essentially everybody is using Claude and it doesn't have a computer. So I'm really excited about giving Claude a computer. And you see the very baby steps of that with the code execution tool, where the model can write code executed on the VM and get the results back. So it can zoom in on images or take a Excel spreadsheet and create amazing data analysis with charts and graphs. And that's just the baby step. What if I had a persistent computer that was always there and it could organize the files in there the way it needed and get the tools set up the way it wanted. And I just think there's a lot of headroom to that scenario. Yeah, and I guess that all ties back into this unhobbling. It gets to it. Exactly. All about the modeling model. That's exactly. Just give, give Claude the tools. Yeah. Well, I'm excited for that future. Thanks so much for this conversation. All right, cool. Thank you.