Model Context Protocol

“And there are a bunch of tools that sales uses that do not have well-supported MCPs or APIs. And I think that's a lot of what made bot so before and after powerful for this group was these were things that they just could not give another AI tool reliably.”

“It's actually really important that these bots have their own computer. And part of it is what I described earlier of there are a bunch of tasks that don't have well-supported MCPs and APIs. We as humans don't do our jobs by MCPs and APIs. Like we use a comput…”

“The key thing is MCP, as much independence as it brings, it's a very runtime fetch of information. You are bound by the latency or you are bound by the search APIs under the hood provided by these themes. So it's a very different paradigm when you rely on High…”

“It can scan AI agent skills for security issues things like prompt injection, data exfiltration, supply chain risk, hidden instructions, basically any malicious patterns and MCP related risks Skill Spectre has your back.”

“Whether you're building agents, they're going to be the super users, or if you're building the software that you expect, either humans, we're going to plug into MCP or other people are going to expose to algorithms and agents, expect them to be super users. Al…”

“It's very possible that there's a, the ceiling on what can be done, you know, with MCPs and like calling out to these server-friendly services has been raised substantially.”

“Things like cloud code things like skills, and most recently cloud design, MCP.”

“It's now a third, roughly as we have been launching it, as we have been launching it partially, I would say, beginning of last year it's now roughly a third, and that combines API that combines credits in product and recently MCP.”

“I think MCP and tools are stupid.”

“We're going from a world in which you have 60 to 80 apps to an agentic AI world where you've got a lot fewer than that. Two, three, four apps where everything else plugs in as an API or an MCP.”

“These models are goal-seeking, right? And they want to finish projects. So Replit just wants, Amelia, you know this, Replit wants to finish. Replit wants to finish, and so does CloudCode if you use it on an own. But when, but if you use Cloud on top of it coun…”

“It is a threat to folks that are building these sort of agent substrates, which is that we, listen, every time we talk about our agents and stuff, a bunch of folks will tweet back or say, Hey, listen, I've got the substrate. I've got the, I've got a way to con…”

“These LLMs are obviously amazing and, but they do lack like the vertical specific context and stuff like that. So you're seeing MCPs, you know, sort of, you have the mob in MCP for designs. You're seeing the idea browser for business building MCP. You're seein…”

“Well, MCP is one stab at solving a very specific problem of agent interactions, which is to selecting the right tools and sort of automating that process. There's a lot more to be solved. It's a drop in the ocean of the overall set of interactions.”

“We're still asking people to understand way too many concepts of like, what is, you know, the difference between like my sandbox environment and production or like connected MCP as myself or others, or how should I store data?”

“And then phase three is largely from January this year, the likes of Cloud Code and Codex and everything that are just ramping up really, really fast. And these are the tools that are using more of the CLI MCP type workload, and we're seeing huge adoption from…”

“I think we're going to do a CLI with skills. So an MCP would be really good if we had more constrained sort of things. However, I think with the number of things that you can do employ, I think the CLI is going to be the way we do it.”

“Because I think what people really want is in my opinion, is really a way of saying, okay, this is actually a virtual employee almost that has, you know, it's on Slack. It has an email. I can actually invite it to me. You can join a meeting, take notes. And yo…”

“When people are buying us for AI use cases, they actually don't really know what they're looking for as much as it was in the past, right? Like we would, yeah, you know, two, three years ago, we go to sell our unified platform to, you know, a classic SaaS comp…”

“The thing we found is, so like MCPs obviously it has been like, it's like really big explosion of, oh, you can go like integrate it with all these different things. But to actually get the integration right and get the right experience, oftentimes we found tha…”

“We've actually given Devon an MCP so they can just go arbitrarily message other Devons and create new Devons, et cetera. But I guess like it somehow creates like a really chaotic world in that sense. And so we, we've still found that most practical use on a da…”

“These things are really good at working with MCP and CLI. Like they're a little even better with CLI.”

“So we're actually gonna have first-class direct integrations with Claude and chat GPT, an MCP server for anything else you might use.”

“How few, even tiny seed startups that are like, bleeding edge, Almost none of them have MCP and CLI, and the ones that do have two users, three users, is not, we are not nearly as far as Twitter would have you believe, right? We're like, oh, everyone's buildin…”

“Actually the initial MCP release includes like a tiny little in memory graph database.”

“MCP inherently has a really strong permission model. Like, all you can do is call the tools. A CLI is a little bit murkier. It's like, can I access the API token? Are you, like, properly sort of, like, re-encrypting the token so it can't, like, exfiltrate it? …”

“So we will always support our MCP insofar as other people are using MCPs, right?”

“And so if we can have our agents properly execute code that calls on CLI deterministically, it's a one time cost, right? Versus constantly having a language model integrate with an MCP over and over and over and paying those like repeated token fees.”

“There are a lot of things where we choose not to use MCP because we want to add more high touch to quality. I think search and agentic find is like the largest instance of that, where we have slack and linear and JIRA search and notion that is not using necess…”

“MCP is a really great way to gain access to tools. Works really well. But you just looked at the trigger UI, for example, there's no trigger protocol. And so those are things we had to build ourselves.”

“What the SDK can do, especially when you're using an agent harness, is it can run arbitrary API requests. So whereas the MCP has all the Zapier trigger and action library predefined out of the box, the SDK can say, hey, I have access to Slack's API docs. I can…”

“MCPs I'm pretty bearish on because the harness forcibly injects all those tokens in the context and I don't really get a say over it. They mess with auto compaction. The agent can forget how to use the tool. There's probably only like, what, three calls in Pla…”

“You're noticing that a lot of people, especially as the models get better, a lot of people throw up their hands when it comes to MCP connectors and they say, I'm not gonna go through like 25 MCP connectors, click auth everywhere, and then like half of them don…”

“By default, most of these agent harnesses, they just have web search baked in, but if you want to actually start linking it up to your tools like Gmail, Calendar, and everything else, which is where the real productivity gains are made, you need to do so, ah, …”

“Anthropic built MCP to basically sit as this translator in between your tools, so that Claude can still just speak English, and your tools can just speak their languages, and this MCP speaks every language, and then just translates your calls from your agent t…”

“My sense is we're making true agents over time the way we think about context and how that context is sort of like shared so that the agent that's orchestrated it actually understands sort of what's behind all these APIs and why and the history will maybe look…”

“We also launched MCPs along with along with this Cloud Agent launch, launch support for Cloud Agent MCPs.”

“But now when you're living in a terminal, for example, and you're using MCPs to talk to LLMs you kinda, you know, the nice to have is actually the UI. The nice to have is the SaaS. The nice to have is going to this website and look pretty. Ultimately what you …”

“What's really powerful about the systems that we're going to go through today is that they exist in the same place where you're building your products and you can get your marketing work and your product work done in the same place and leverage skills and MCPs…”

“Like you could have one focus on research and using MCPs. Those can fill up like context pretty quickly. So sometimes it helps to keep that in like a split terminal and then do other work side by side. So you can like save research, save files here that you're…”

“In terms of MCPs, like I used to have like tons of them set up. And what I realized was that simple's often better. I have perplexity. I have fire crawl. I have playwright. Those are like my main ones”

“If you need to, you can use MCPs on the fly. You don't have to restart, unlike, unlike Codex or Cloud Code, where you actually have to restart the whole thing. I think it's way more elegant, and also scales way better.”

“Because my premise is MCPs are crap. Doesn't really scale. People build like all kinds of weird search things around it. But you know what scales? CLIs. Agents know Unix.”

“Now you have even cloud skills and Claude, so NA-TEN MCP that can build the NA-TEN automation for you. You know, you can kind of brainstorm with Claude and Claude code and they'll be like, great. Can you now use the MCP of NA-TEN and just build it? And it just…”

“Don't over obsess. Obsess on MCP skills, et cetera. I'm not saying don't get into these. I'm not saying don't read about them. I'm not saying don't use them, but I can almost guarantee you these things are not the reason why your product isn't working, right? …”

“There's actually a lot of value in having like multi-turn conversations from like the orchestrator or the assistant to like the sub agent, whereas like, you know, a tool call is basically just like one RPC.”

“I think of like the MCP and tool usage as being like the interface to all of our conventional imperative systems, not at the AI space.”

“I would say that this stuff like barely works in fairness right now.”

“I think MCP has been a huge driver of context engineering.”

“And I think for us, for me personally, like in my dynamic agent configs, I'm moving more toward API calls. And something a little bit more once I kind of maybe been able to prototype with an MCP server and figure out how I'm going to use this I think then you …”

“And then like, you had this like inflection point around April with like Sam Altman and Satya and Sundar and all posting about like MCP and that they're going to adopt MCP at Microsoft, at Google. At OpenAI and that was really like the big inflection point.”

“We just launched like in our end of like November, then the recent iteration of the protocol, finally, like the next bigger improvement to the protocol, which is like long running tasks to really allow for like, you know, deep research type of task and like, y…”

“At least on the protocol side, like a defector standards. And I don't think any of the other protocols, it feels like they're not just there yet, but of course, if they get there, then we are like super open as long as they're like complementary to what's, wha…”

“In the first iteration of our authentication respect, we combined them together into the MCP server... And if you combine these into the same server, you just can't do this anymore. And so all we needed to do is like, okay, we are a resource server. The MCP se…”

“But if the client and the server don't know each other, we just don't have a good solution for now.”

“MCP should also enable agents, and agents are inherently somewhat stateful, and there's some form of, like, long-term communication going between, like, the client and the server, and so we always looked for something like that.”

“We got wrong that. We made a lot of things optional for the clients to do. Like you can, the client can connect and open this return stream from the server, but it doesn't have to. And the reality is, is no client does it because it's optional. And so a lot of…”

“And we know from like some companies like the Googles of the world, the Microsoft of the world, they're doing MCP at a scale that I can't tell you the numbers, but it's like in the million of requests.”

“And so that's why we run MCP as like a really traditional open source project with like a core maintainer group of like eight people that basically decide everything. And then like input from everybody else. Like we get input and people can make suggestions an…”

“MCP is a protocol between the AI application and like servers, right? So the model is actually technically not involved in MCP.”