Welcome to bots.so

"An army of robots is already here. It's very cheaply available. The bottleneck is just figuring out intelligent and interesting things to do to them."

That quote is on our home page for a reason. The compute is here. The APIs are multiplying. The providers are launching faster than anyone can track. The bottleneck isn’t access anymore — it’s discovery.

Why bots.so

We named this project bots.so because we’re not just thinking about LLMs. We’re thinking about every kind of bot — language models, image generators, voice synthesizers, robotics APIs, autonomous agents. The whole spectrum.

Right now, if you want to find an inference provider, you Google around, maybe check a few familiar names, and go with whatever you find first. You never discover the dozens of other providers that might be cheaper, faster, closer to your users, or better suited to your workload.

We think that’s broken. And it’s only going to get worse as the market grows.

Starting with LLMs

We’re focused on LLM inference providers to start. It’s the most active segment of the market — new providers are launching constantly, existing ones are adding models weekly, and pricing shifts all the time.

We’re tracking 30+ providers today. Most developers only know about 3 or 4 of them.

AI from everywhere

We noticed something early on: compute is centralised. The same handful of US-based hyperscalers show up in every “best AI API” listicle, while providers from Europe, Asia, the Middle East, and beyond get zero visibility.

We want to change that. bots.so actively seeks out providers from places that are new to the compute game — whether that’s a GPU cloud in Iceland running on geothermal energy, an inference startup in India offering frontier models at local pricing, or a sovereign AI platform in Europe built for GDPR compliance from day one.

Diversity of compute matters. Different regions bring different strengths: energy costs, regulatory alignment, geographic proximity to your users, and hardware choices you won’t find elsewhere. We want the index to reflect the full global landscape, not just Silicon Valley.

If you’re a provider from a region we don’t have listed — or you’re planning to start offering inference and compute — reach out to us. We want to know about you.

How we find providers

We spend our own money running multiple state-of-the-art deep research tools to continuously scan the internet for new compute providers. When a new provider launches or an existing one adds new model APIs, we find it, verify it, and add it to the index.

No provider pays to be listed. No rankings are bought. We find them, we list them, and you discover them. (For full details on how we operate, see our terms of service.)

That said, we’re a small team and we need to keep the lights on. We may include affiliate links for some providers — if you sign up through one, it helps us keep building bots.so at no extra cost to you. We’re also actively hunting for discount codes from providers to pass on to the bots.so audience. If you’re an inference provider and want to offer a discount code for our users, get in touch — we’d love to feature it.

One email, everything that matters

We wanted to keep it simple: one email that tells you what’s new. New providers that just launched. New models that just went live on existing providers. New APIs that are ready to call.

That’s it. No noise, no marketing, no 47-paragraph newsletter. Just the signal.

Where this is going

LLMs are the starting point, not the destination. We’re expanding into other types of ML and AI — image generation, speech-to-text, text-to-speech, embedding models, and eventually robotics and autonomous agent APIs.

The vision is straightforward: we want you to be able to come here and find 200+ compute providers for whatever you’re building — whether that’s a chatbot, an image pipeline, a voice assistant, or a home robot.

The compute is here. The army of robots is here. We’re building the index so you can find them.

Explore the providers and see what you’ve been missing.