Apply now
If you're running production LLM workloads and want microsecond-overhead routing with founder-level support, send a brief note about what you're building. You'll work directly with the person who wrote the hot path:
The same C++ engine that powers llmbridge, plus everything you'd otherwise have to build: real-time routing intelligence, observability, team management, and managed infrastructure. Managed by us. Used by you.
Multi-region deployment, automatic failover, and an uptime SLA. You point your client at our endpoint; we operate the gateway and handle the rest.
A real-time "shadow order book" measures every major provider — actual latency, throughput, and error rates — and routes around degradations before status pages update. The same open-weights model can be priced up to ~10× apart across providers (≈10× for Llama-3.3-70B right now); routing captures that gap. We estimate 15–40% lower effective cost — measured per customer as design partners go live.
One OpenAI-compatible API across OpenAI, Anthropic, Google, Bedrock, Together, Fireworks, Groq, and more. Switch providers without touching client code.
Per-request OpenTelemetry tracing. Token-cost attribution by user, team, and project. Budgets, forecasting, and audit logs. Prompt content is never logged.
You work with the engineering team directly — same-day response, custom integrations welcome.
OpenAI-compatible surface. Change your base URL and API key; existing LiteLLM / OpenRouter / Portkey code works. Most teams migrate in under a day.
Tokens are at-cost passthrough to the LLM providers — we never mark them up like OpenRouter; you pay exact provider rates. Our fee is a transparent platform + per-request charge for the routing service: a stated line item, like an execution commission — never a hidden margin on your tokens.
| Plan (at GA) | Price | Best for |
|---|---|---|
| Team | $149/mo + usage | 1M requests/mo included, then $50 / 1M requests |
| Business | $499/mo + usage | 10M requests/mo included, then $25 / 1M requests |
| Enterprise | from $2,500/mo | committed volume, custom volume rates, self-host in your VPC / on-prem |
"Usage" is a flat per-request routing fee beyond your plan's included quota — not a token markup. Tokens are always billed at the provider's exact rate; our fee is a separate, stated line item for the routing service (an execution commission, not a marked-up price). Every line on your invoice is computed from request logs — nothing estimated.
The routing opportunity, made concrete — and the catch. The same open-weights model is sold by 15+ providers, but not as the same product: they differ in quantization (fp8 vs fp16 — i.e. output quality), context limit, latency, and uptime. As of 07/28/2026, Llama-3.3-70B input runs about $0.10/1M on the cheapest provider versus ~$1.04 on the priciest — up to ~10× — and the cheapest is usually the quantized / smaller-context one. So routing isn't "pick cheapest"; it's pick the cheapest that meets your quality, latency, and error budget. Live price and specs, side by side:
| Provider | Input $/M | Output $/M | Quant | Context | Uptime |
|---|---|---|---|---|---|
| Fetching live prices from OpenRouter… | |||||
Fetched live from OpenRouter's public API when this page loads — figures are OpenRouter's per-provider rates (mostly provider pass-through), not a Kottos quote.
llmbridge is the fast gateway core — format translation, microsecond overhead, run-it-yourself. The hosted gateway adds the routing intelligence (the shadow order book), managed infrastructure, observability, team management, and enterprise features. llmbridge is free; the hosted gateway is a subscription plus a transparent per-request routing fee (never a token markup).
We continuously probe every supported provider, measuring real-world latency, throughput, and error rates per model. On each request we route to the provider that's actually fastest and cheapest right now for your model and constraints — not the lowest list price. The same model can be priced up to ~10× apart across providers (≈10× for Llama-3.3-70B today), so the headroom is real; our 15–40% figure is an estimate from that dispersion, and we'll publish measured savings as design partners come online.
No. We never train on customer requests or responses. We collect anonymized latency metadata to improve routing; prompt and completion content is never logged or used for anything beyond serving the request.
Yes, on Enterprise: deploy the Kottos gateway as a Docker container in your own VPC, so prompts and provider keys never leave your network and it sits next to your app for full sub-millisecond latency. It connects to the Kottos control plane for real-time routing intelligence and usage metering — the routing "order book" itself stays managed by us. The open-source llmbridge core (translation + proxy) is separately self-hostable under Apache 2.0, no account needed.
If you're running production LLM workloads and want microsecond-overhead routing with founder-level support, send a brief note about what you're building. You'll work directly with the person who wrote the hot path: