The Infra Pod - Building a successful infra product between all the AI apps and model providers (chat with Louis from OpenRouter)

Episode Date: March 9, 2026

Tim (Essence VC) and Ian (Keycard) interviewed Louis Vichy, co-founder of OpenRouter, about why he built OpenRouter to de-risk AI app development (end-user pays LLM costs), how it scaled to processing ~5–6T tokens/week, and what OpenRouter is today: a reliable inference routing/control layer across ~60 providers with consolidated billing and reduced vendor lock-in. Louis explains why teams adopt OpenRouter (constant new model integrations, pricing/billing, differing API shapes), how routing focuses on practical heuristics (fallbacks, cost, throughput, latency), and how reliability is achieved via provider failover (e.g., alternate endpoints like Vertex/Bedrock). They discuss agent trends (longer-running agents, small models for routing/classification with specialized downstream models), possible memory support, developer conveniences (e.g., PDF parsing), and enterprise features (security/compliance guardrails, presets). The episode ends with links to OpenRouter chat/rankings pages and hiring for high-agency TypeScript-focused engineers.00:00 Welcome & Meet Louis (OpenRouter Co‑Founder)00:27 Origin Story: De‑Risking AI App Costs (Hackathon Lessons)01:35 First Big Feature: End‑User Pays for Tokens (Sign in with OpenRouter)02:34 From Routing to Rankings: Scaling to Trillions of Tokens03:42 What OpenRouter Is Today: Reliable Inference Across 60+ Providers05:55 Why Teams Adopt It: Avoiding Model API Churn, Billing, and Vendor Lock‑In08:37 Winning Strategy: Don’t Build a “Magic Router”—Optimize Cost/Latency/Throughput18:58 From Chat to RAG + Memory: Building Persistent Agent Context20:37 Developer Bells & Whistles: Auto PDF Parsing and More21:11 Enterprise Readiness: Compliance, Security Guardrails & Model Presets22:22 Customer Growth at Warp Speed in the AI Era23:03 Spicy Future!

Welcome to the InfraPod. This is a 10 from S&VC. And Ian, let's go. Hey, this is Ian. Lover of all agents, couldn't be more excited to talk to Lewis Vichy today, co-founder of OpenRouter.

Lewis, what got you started on Open Router? And why did you dive in to start in the company? What was the crazy idea or insight you had that got you all started on this journey? Oh, starting from the bottom.

Essentially, one of the thought that we had was building with any, new technology would incur a lot of risk. And this actually was inspired by a lot of the hackathon journey I went through. So I went to a lot of hackathon early on in my day back in college. And every time I heard story about a kid who used the MongoDB free plan, and then they use this so much.

And then they're publishing an app, right, with the production pipeline. And then their free tier, you know, like kind of expire. And now they're like on the hook for like $70,000 for MongoDB. and you'll be like, oh my God, how are they going to handle that? And I would imagine AI, this is early 2020. I imagine AI would have the exact same issues, meaning people are going to build crazy stuff with it, and people are going to consume a ton of token.

And it would be so risky to build with a proprietary AI models unless you have to do some local models, right? That would be much more sustainable, meaning the models running locally, and you wouldn't get charged much for it. But then the problem is local models sucks, right, at the time. None of them were good. And so the first thing the open router ship was really a way to derisking developer entirely when they built AI apps. We built basically this feature that allows you to do sign in with open routers, where the user of the app paid the LM directly, right?

So essentially, user flow where you would create an open router account, and then we will mint a key for you. And that key will allow developer to charge the end user directly. So a lot of apps that start with us, that built with us early on, they would consume billion of tokens and the deaf paid nothing because the end user paid directly. So that is really, like the thought process of the first feature. And then once we got the feature out, right, when the gate is open where a lot of developers built, right, and they can build crazy, crazy agent and loops with us with practically like zero risk. I mean, there are some risks, right? But it's not like a financial ruin. We start thinking about how can we scale the model pipeline, the model evaluation, the model.

How can we make the model more competitive, right? How can we make all the model labs more competitive? And that is the origin story for the model ranking page, where we are now, if you look at OpenRuhrer AI slash rankings, you can see we are doing more than north of six trillion token a week. And that has been a crazy journey. The first year, when I showed to my dad, we, were processing about roughly 4 billion token in a week. The second year, I showed it to HF0, one of our precede investor, accelerator. We were processing about 150 billion token in a

week. And last year and essentially growing now, 5 trillion to 6 trillion token in a week. So there's clearly a growth curve of the thing that we are processing on over browsers. It kind of proof like there's a PMF there, product market it, and also proof that more and more models are being more useful and being more competitive, right? So essentially that is like the, from the initial

journey of open-rador so far. And where are you today? Like, if you were to sit down and say, okay, obviously open-rroters try to head, like many things, very humble but insightful beginnings. You know, many people who are in building agents have heard of open-router one way or the other. Like, could you help people understand

what OpenRotter is today? and how that vision has grown over time. I'll say today we aspire and we hope we are the most reliable way to source intelligent, to source essentially inference across 60 different providers. When you need inference, you go to open routers, right? And we hope that what we're providing is the add-on on top of the models, labs, and the provider themselves, right, to help the developer built with essentially derisking them. They risk them from

Beside financial room, also vendor lock-in, right? A lot of time, we also help developer like building relationship directly with provider too as needed, right? Because a lot of times, they would just then bring the key from the provider over to us so that they can manage all the AI

span on open routers, right? So essentially we are now slowly evolving from the best place to get inference into like the control plane for all of your inference need. In France, essentially, you call an alarm, you call chat GPT and so on, to get, you know, some token back. And so I think, like, I'm a developer, Ian's developer.

We're all very much all using OpenRruder Summer, which is like what a fun part of it, right? And it's like, it's out there so, so widely. But if you ask me even like maybe two years ago, there will be an infra company just be routing. Not just be, but like there will be centered around routing to variety LMs. it will be a little bit unthinkable, like, okay, how big of a problem that will be, why we need a layer that exists? Can you talk maybe a bit more like from an infralayer? Why is everybody adopting you?

Like, what were you thinking about? Is this too much complexity when I'm dealing with a variety of model plus and minus providers? Or there's some, even like some lot more nuances. Maybe unpack like a complexity here. Why is it such a hard thing that people don't want to build themselves to rather use open router today? I think, well, to your point, right, so many people early on for the first, so the company been around for three years now, I think, and the early journey, people would look at us and

the prosumer, the enthusiasts would be the most stickiest and we stuck to them, right? So they would be the highest advocate for us. And we understand the pattern. The pattern is that model labs, I mean, the bet is, they're going to be better model in the future, right? That was the core bet for the first year and a half. and with that bet every time there's new model

coming up on Tuesday I don't know why but they always ship it on Tuesday your company you're going to spend an hour with all the engineers scrambling to support that new model if you don't

your customer will be like hey I want Sonnet 4.5 or I want the latest GBT model right and then you have to add all of those code and now you also have to add tracking for pricing

billing and all that stuff right and even worse when you have to move to a whole different API shapes, like moving from open AI to anthropic, totally different shapes, right? And I think after people like onboarding about 10 models, they were like, yeah, this is enough, you know, like, I'm done. It's very, very, you know, like labor, it's basic labor. And so we have built all the Army system

to track provider. We work on a spec to help provider, like, you know, communicate with us, you know, the changes of all these endpoints over the past years or so. And so we have, have the infrastructure to keep track all that together with pricing as well. And so I think that is the main reason why, you know, people are like, yeah, let's just stick to open router for now. And then, you know, probably easy for the entire team. And we basically save you like, you know, an hour almost every week, right, on engineering time, doing some media work. And then on top of that, over time, we also add, you know, all the bell and whistle feature, right? Like, we add, you know, plug-ins. I mean, over time, we actually

built feature for both developers and also their manager, right? Because, like, if you have AI spend across a bunch of vendors, it's just like a nightmare, especially for a small team. And your finance, your CFO is not going to love it. But the nice thing about open router is you manage one bill and you have, you know, essentially consolidating usage across 60 different providers, right? And your CFO would be decently happy, I would say, to our business. boarding us. Yeah, that's actually, that's amazing because as you mentioned, like, model,

number of model is just increasing. Yes. And I think we are also in this huge debate between like Uber models versus even smaller models and all this kind of stuff. Actually, maybe it's a fundamental question. Like, I remember when I was looking at a variety of people building LM apps, there actually has been a variety of libraries out there to try to abstract away, calling different models as well. And so I think probably the most common pattern I've seen is like some

GitHub library project that lets you able to extract away different models and stuff. But you guys are totally not just that way. There's a hosting environment and stuff like that. I'm sure. Maybe the way, right question here is like, why do you guys won?

What got you guys to be able to be almost like the number one adopted product, when there's actually a lot of spring of lots of things happening a member at the... I think this applied to almost any company out there as well, by the way. So when we were building our open router, there were also a lot of competitors, right? But a lot of them were focused on the wrong problem, meaning they're trying to solve for the magic routers. A lot of early model did not make the bet that there will be always better model in the future. They make the bet that the current model are good enough. So let's try and find a way

to find the best model of the current model. But we focus heavily on that is going to be better modern futures, like focusing on measuring what we have on the current model, but automate the pipeline to get the next model and build the relationship, right, to get as much capacity we can get on the next top tier model, right? And build a very strong relationship with all the AI labs, because I think that's like the awesome part of open routers.

We stood at the standard of all the AIM. models lab and the app developer and bridging the gap, right, across apps and all the AI models. And going back to the point of like doing the thing that is like, we didn't try to build like this magic router. The router is very, it is a very, the heuristic is decently simple. It's essentially falling back. And we're routing based on either pricing, throughput, or latency, right? These are stuff that you can track.

These are stuff that you can track. You can realistically track. You don't have to eval the model to see if it's smart or not. And these, I think, are the core insight that we have done over the past year and a half where let's focus on that first and build a smart model later. Because the core need of a lot of users, though, is still, can we optimize the cost?

I actually just need models. It's extremely fast. The faster deployment, right? And I think just listen to your customer to see what they actually need at the moment. Apparently not a lot of people need this magic router to the wrap between the model. And I think this applies almost to every other technology

the company out there too, is focus on what can be realistically do and focus on the customer first. And naturally, a customer would stick to you and ease you more and naturally spread you like Wi-fires. And I think that's a pit of success for open-radorers so far. Yeah. I'm kind of curious, as models evolve, as new model types, you know, different multimodal models, we've got all great for speech, robots, great text, different reasoning, coding, whatever. I'm curious, do you think we enter a world of like, there's like a long tail of specialized models and the use of the model is like hyper-task specific?

So like today, let's be honest, like today you have an agent, let's say cloud code, right? And that agent is very specific to coding. And Opus 4-5, Sonnet and like the coding-specific like models are like, that's its jam. And what we've learned also is that coding is increasingly becoming a solved task. and I'm curious, do you see a world where we need like a generalized task model routing layer that's intelligent to be like, hey, the user wanted me to do this? Like, what do I do? Or do you think that there's simple heuristics and simple agents where we end up?

Like, what's the complexity of like what tasks agent, at high level when I say agent, like the top level agent for like a dummy, like a chat GPT or whatever? Like what's the level of complexity on the routing and the task and how that's interrelated with sort of the prompt that. that the user is asking. My hot tech, this might be a hot take. We love hot takes. Yeah, I mean, we're here for them. My hot tech is possible. You just need a very small model to make two core decisions. It could be a bird.

You might just need a very small model to do like some kind of edge routing on like the kind of classification decision. Decision workflow. And then downstream, sure, can be a mixture of huge giant model. Because if you look at ClarkCode, clock code essentially just clause sonnet, right? and claw opus. The model is opus. It's very generic.

The thing that make claw code is the harness, the agent harness, which is like, it's a library. I'm sorry, it's a binary, running on bun, probably written in Thai script. And that can certainly be replicated by a lot of people. And there's a lot of people trying to replicate what Clock code is doing already, right, in the open source space, like open hand.

There's also like other alternative closed source solution like Devon as well, right? like in from cognition. So I would say my hot tech is the base decision layer can be a very small model. But it will fan out. It will fan out to like more space of model. Gotcha. And I'm curious as open router evolves, as you think about the space and think what's happening, is that sort of the open router vision is that we're going to be right there along the way

and get smarter over time? Or how do you think open router revolves as the space evolves? And I'm also curiously, how do you think the complexity of these agents evolve as well? That's another word to ask a previous question, but in a different lens. I think the agent is going to run for longer, for sure. People want to try to make them run longer, running overnight even. What am I dream right now? I'm trying to build this thing as like my personal goal.

It's like when I go to sleep, I have like a farm of agent just, you know, building stuff for me, right? We already have that, but like, that would be fun to have them just keep on running. I mean, I don't know, but basically my current habit every day is before I go to sleep, I just spit up like 10 different taps of clock code and just say, hey, refractory my code base for me. And then in the morning, you know, hopefully it will come up with something. I will review their PRs,

which is, but I would like to automate that even more. I think the current baseline, all the product is still like, let's make sure we are the most reliable way to source inference. So let's make sure that the base layer is extremely reliable, scaling of infra, making sure that we have enough, like all the Q are

you know, acting, not knocking, all the description are, you know, like, flowing. So that we can serve the baseline. Open router essentially become like the utility layer to get the model, right, to get like the inference, which is like how we can call tooling and so on. And then

the, and then let the developer build it out. The developer will build out the agent harness, the infinite loop, you know, the temporal workflow that's running on top open routers, right? That's like of Hawaii, right now, we are thinking about the positioning of open riders, right, to become a very reliable infrastructure for, like, all the AI agents.

And talking about reliability and stuff, I think looking at your website, obviously, performance reliability is almost like the forefront. And then there's just billing and stuff like that. You know, when I talk to friends, even two years ago or even now, like open AI being down and thought with being down, it's almost like a given every day almost like, ah, my life stops now. I can't Cloud. I can FiveCo because cloud is not working. It seems like it's so common. And I guess since you're in the middle, right? You know, if that topic goes out, you can't really do anything, I assume. I guess, oh yeah, so I guess the question is like what is, what are some of the things you're trying to do to really, maybe even technical challenges that are surprisingly

that people don't know of trying to solve this problem is give us a little bit of like a taste of what are actually some interesting problems you solving for customers? The customer may not even know about. Well, essentially, let's say Anthropics is down, right? But there's only Anthropic first-party API. There's still a vertex that's serving the anthropic model, and also there's Bedrock that's serving that Anthropic model, right? So essentially, the router would follow you back on the same model across those endpoints.

And essentially, in the back of the room, we would talk with, you know, like, Google and Bedrock to secure as much capacity we can, right, to serve when there's downtime with your anthropic first party. So if you went to any of the, like some of the anthropic model, even opening eye model, right, in the uptime chart, there are time when the upstream endpoint went down, but our uptime is still, you know, like in the 99. And it's all because we have this mechanism that we have engineered and design so that you can fallback. And these fullback mechanics is very model-specific, right? Each model has a bunch of providers.

And that's like one of the challenges. And then there's all the challenge of it across these endpoints. So for a given model, there are, let's say, five different provider. A model and a provider within OpenDLINCO, we call them an endpoint. And so we have five endpoints. Each of these fine endpoint, though, might actually have different ways of service. the model. Some might have tool calling, some might have structured output, some might have none. So these are the things that we would then have to look at your request and figure out, okay,

which one is the best endpoint to serve your request, even given for a single model. I don't know if it's well known, but we have like a parameter in the API that allows you to specify the max pricing that you're going to pay for this request. And so we would have to then, you know, like measuring how expensive your API, your API code would be to essentially estimate that and route you to the right endpoint, for example, so on and so forth, right? So there are so many small little things that really our developers have worked with us and kind of expressed to us and we just build them out.

Yeah. Yeah, that's actually really super interesting because there's so many little problems like that. Just using your normal models, you think, or just I'm just calling a model, but behind it saying there's a lot of things that need to happen. I actually want to maybe double down in some other aspects of things. Because I think today, when you look at models, used to just be chat GBT sending a prompt. But today is a lot more about rags.

And even beyond now is a memory, right? Like actually, how do people consume and how to remember the prior, either it's a multi-turn state kind of thing, or there's going to be a way to remember prior states in some level. This open router has a role of memory in spiritual speed. If people want to have persistence or even memory across agents, calling a model, is it just like, okay, that's the last one you call go there? Or do you have even some aspects of like, I want to able to save memory and go across different endpoints?

We are definitely very interested and we are talking actively with customer, right? I think this is the part where we always, we want to have deep custom empathy to really understand like how can we serve memory. without entrenching, right? Because a lot of times when we build memory, memory is also another thing where it's very, if we do it bad, we will lose trust with the end users, right? So I think it's crucial that we designed with,

but to your question about, are we interested in like kind of expanding the API to be slowly, you know, like improving the experience when they're building up Asian? I would say the answer definitely yes, right? Because I think more and more people are building out Asian. And so any feature that they're risking developers and allowing them to build better workflow, better agents, we are all for it. Yeah. A lot of model providers, right, doesn't really parse PDF, right?

A lot of times, they don't. We can act them to parse the PDF, but a lot of times they are just serving the basic LM models. There's no PDF parser. And so we would then add a layer on top of our API to. always parse for PDF. And so when you send a request with a PDF, it is all my employee, like, parse for you. Yeah. So like those are the bell and whistle and stuff that we added to the API to help the developers. I mean, there are also other stuff that we build as well. I mean, as we grow and as we're getting adopted by bigger customer, we now also

look into compliance, looking into like security posture, right? There's like, we have a new feature coming out soon called Godbrow, which allows to actually controlling, right, how your model, for your entire organization, right? Because when you're a Cecil, you kind of want to ensure that your organization is using the right models, or using a model that has data center in the US, or have like software compliant, hippocomplying, or even have like, you know, like you want to use endpoint, right?

Endpoint meaning a model served by a certain provider that meet your, you know, like maybe data retention criteria like ZDR. So we also have all this feature built out for, you know, enterprise to serve their need. And even for enterprise customization as well, we have this feature called preset. So with a preset, you basically make an alias for

a model with a custom display name, with a system prompt that you can add to it, right? So now you have this nice little custom model that your company can use across your organization. These are the kind of thing that we are incrementally building out as our customer

grow because interestingly, the journey of open out is a lot of our customer become more sure, right? They come in, they build a small little company, and then either they grow to be bigger or they bring us to their actual work, to their boss.

Yeah, so it's a crazy journey and it's very fast. Like, the iteration is so fast that a lot of them have graduate from like, you know, like, from literally a one-person team to now like a 25-person team something. That's like, yeah,

that's like into like a series company. So it is fascinating to see your customer growing just as fast as you, I would say. Yeah, AI is nuts. Well, it is not. Let's jump to our favorite section, what's called a spicy future. This is where you give us, what's your spicy hot take around AI or infra that most people don't believe in yet? Yeah, I was thinking about that, but then, you know, as we speaking, I'm like, hmm, is that like a good idea? Is that a fun idea, right? Okay, the question is, what is something that you think is true,

but everyone's thing is not true right now? But focus on, like, AI infrastructure. Or just AI in general. Yeah, it doesn't happen. And AI in general. Yeah, yeah. I know. The problem is, you know, like, so one thing with me, though, right,

is this is like a separate topic. I'm just like brainstorming collaboratively. I have a lot of these thought about half a year ago, but as a company you grow, I have to then condense a thought into a spec, you know, RPD and thing, right? And then present it to, you know, my team so that we can actually internalizing it. And then, God damn it, Anthropics released a bunch of freaking paper and article about certain things, like long running agent and skills and tool calling. And everyone's, oh my God, yes, we got to do it.

I'm like, fucking it. I have my head in there, you know, somewhere. I just went down to like a PDRD or something. Yeah. So it's very hard to get a tick now, at least from my head right now, where people don't agree because it's like it's not all news. Unfortunately. That might be a hot take on its own, right?

Which is like people, anyone's believing that they have, you know, revolutionary thing, a spec or something new, is not new in AI quickly, right? Just the level of like changing. And the state of art model that we have can only stay state-ed-art like a week now. Yeah, very much a week or two weeks. Crazy or stupid, like low, ephemeral these models can get, which plays in your favor. So maybe you can talk anything in this line of wording at all, actually, right?

It doesn't have to be a very, very specific product or project or technique. It could be just like everybody still believes, you know, open AI would just win over, right? Or anthropic and everybody. and everybody. So that was the bet, though, but that was the bet that we made originally with open routers. But I still feel like today may not be still universally believed, right? So it doesn't have to be like, no one believes.

It can be like majority don't believe. Another part is I also don't want to be too forward-leaning to be favoritism or like criticizing one partner because they are our partners. Yeah, they are all your friends. I know. Yeah, they are. Yeah. We don't want to do that.

We don't want to do that. You don't want to do that. You're very neutral. party, I understand. Yes. So it's also tough for me to get an actual hot tick, right? Because like, for example, some guy from anthropic and literally came out as like, nah, two calling for opening eyes like a mid implementation.

Then I'd be like, yeah, I guess. You got the right to say it because they're compared it, right? They can slingshot each other. Yeah. Yeah. Maybe, so yeah, maybe another way to do this could be like, what are something that you see are starting to happen but hasn't happened everywhere yet? But so one, I'm starting to see a lot of companies do a small company experimenting with this.

So linear just released their linear agent, right? So now you can go on Slack and tell linear, hey, let's make a ticket for me. That is not open in public. Fern also have a chatbot now where you can tack Fern and it will improve your doc for you. And I think there will be more and more, are you familiar with like Devon from cognition, the AI software engineer? There will be more Devon for X. Devon for Excel spreadsheet

Devon for update my docs Devon for HR that will live either in your Slack in your team message in your bunch of your your chat interface essentially wherever you will consume or chat with other people I think

there will be more of that So I think seeing more like the Devon or people, somebody even use the word cursor for X or whatever that. That usually implies there's going to be one dominant player that does that. Because encoding,

even though there's many, many other players, cursor has been almost like the number one player everybody uses. Cursor is awesome, but I think clock code is like gaining huge traction because if you look at the way that atropics has positioned in clock code, right?

The clock code agent can be decouple. And as mentioned before, the clock code is just a harness. You can take clock code and run it on a cloud container and then use it to run like an Excel spreadsheet or to running your entire finance team or like to run a bunch of interesting stuff

agentically, you know, in a background, right? You can now implement it with claw code. My prediction is clock code will be rebrand as claw. So it will be just a claw agent. It's a base agent. And then you can deploy claw agent to almost everywhere. And I mean... A co-worker thing they launched, right?

Oh, yes, they did. They launched that, which I'm saying, right? Any hot tech is now. They haven't really retired code. I think they just added a coworker that does think in your computer only. But I think if you're being going further, like, it's not just your CLI for coding and not just doing stuff in your computer. There will be more, I guess, environments that runs in other data and other apps running. Pretty much so, yeah.

And everyone can actually build that. Like, not just Anthropic. Everyone would actually leverage what Anthropics built to build their own version of it. I would say. Because the main reason I've had a lot of friends try out co-work, and the main

response has been, it doesn't seem like it will help me in my workflow. But imagine if they can build co-work with their own workflow, and if that building

process is very simple, then I think more and more people would do it, right? Maybe even spicier, because, like, Claude, obviously,

cumor-anthropic, will just support Clod, you know? Cursor of the world are independent toolings, they support all models, right? And so a lot of these products are using yours. But you have the advantage. I guess I'm curious for you, do you see or believe in the future,

everybody doing finance will all use one or two models only, and that's it, you know? Or do you think, like, the tools will be the most valuable in some sense, and because the models are all different, basically everybody will use different models, and there will be like a winner-take-all in all. all the tools models, you know? Because I think that open route would be the most interesting layer because no one, no of what single model went out. Everybody, there's always a need for a variety

in ongoing improvement models. Do you think there's going to be a monopoly model every single sector here or not? My take is that even if there's a monopoly is very easy to disrupt it. If there's a monopoly, a new model lab will just come out of China and just disrupt the whole thing.

So there's no monopoly. I mean, if you go popping for like a day, That doesn't count. Yeah, not a job. But yeah, essentially. It's not a belief. That's just like a blip. It's a blip, yeah.

Yeah. So in my mind is really, this really strengthened, right, open route, like kind of positioning as like, with a layer is a reliable layer to like get access to all this model, right? And so the harness is separate from the model, essentially. And I think the best harness should be able to reliably extract, you know, like a workflow from any model, to be honest, for the most part. Yeah. And then, sure, if the model is dumb, switch model, open router is there.

Yeah, yeah, that's amazing. Okay, I think that is a really interesting take here. So for folks that, I mean, I'm sure everybody has heard an open router, but for the random people that hasn't, or want to check out more or even want to try out some of the models, where would they go to find OpenRouter or find you? Let's go to OpenRouter AI. We also on X, and OpenRodder AI slash chat.

If you want to try out some of the model, you will have to sign up for an account, though. But there's a plenty of model for, you can re-offer for free, together with some of our partners. You can also go to rankings. OpenRuards AI slash ranking is the most interesting page. A lot of people have, you know, like, retweet this page. It shows the actual token being processed by open routers. And you'll see it's grouped by models. But as you scroll down that page, you'll see it's grouped by some use cases. like programming, like marketing,

like, you know, copyrighting, so on and so forth, right? We also have, like, model proved by languages, and eventually we'll have some, I think we'll have a ranking based on throughput as well. Yeah. So it's the best way. Precisely, yeah. It's a Google trend for AI.

You take, you, like, a lot of company would have literally a screen that's showing that ranking page just so they can track, you know, themselves versus other people, which is a very flattery thing. Also, an operator, we are looking for extremely high agency engineer. So if you're looking for a place where you can basically use any AI tooling, you can use any agent out there to build amazing stuff and automated workflow,

let us know. You can ping me directly too, Lewis and Open Order AI. Amazing. Yeah, so you're a stack engineer with a ton of agency, You want to try any AI Dev tool out there to be in the hotness of everything. Open route is the place to consider, right? And make sure that we love engineer with a good taste. And especially if we use TypeScript because we use TypeScript for almost everything. Okay.

Well, Taste and TypeScript, not everybody will put them those two words together sometimes. That's the reason why Anthropic Bot Bunn, right, and a bunch of the TypeScript ecosystem tooling has, you know, like, even though it's still very new, there are certainly strong tastes. in a sense of like, when you use TypeScript, why do you need TypeScript? And like, I mean, the whole why TypeScript exists, right? It's, I think, a fascinating story by itself, I would say. Yeah.

Right. It helped with the type safety of JavaScript, which is already a crazy language. But it's basically ruining the web, right? So I think it's like, I think it's a good heuristic to find people who are, you know, like truly lean into, you know, a certain ecosystem, I would say. Super cool. Well, hey, thanks for your time. Thanks so much.

I think we have asked some more questions, but just based on time, I want to stop here. Yes, thanks for being on our pod, and I'm sure I've been a ton for OpenRouter.