One API, Every Model
Unified access to top-tier models from Anthropic, OpenAI, Mistral, and more. Whether you are using Claude Opus 4.8 or a specialized open-source model, Kilo Gateway routes your request intelligently.
Kilo Gateway
Route LLM requests to any provider, anywhere. Use Kilo Gateway to power your own apps, or bring your own keys to access models without a subscription.
1import { streamText } from 'ai'2import { createOpenAI } from '@ai-sdk/openai'34const kilo = createOpenAI({5 baseURL: 'https://api.kilo.ai/api/gateway',6 apiKey: process.env.KILO_API_KEY7})89const result = streamText({10 model: kilo.chat('anthropic/claude-opus-4.8'),11 prompt: 'Why is the sky blue?'12})
Why Kilo Gateway
One endpoint, every provider, and full control over how requests get routed and paid for.
Unified access to top-tier models from Anthropic, OpenAI, Mistral, and more. Whether you are using Claude Opus 4.8 or a specialized open-source model, Kilo Gateway routes your request intelligently.
Don't want another subscription? No problem. Plug in your existing API keys (from providers like Z.ai or MiniMax) and use Kilo Gateway as your centralized infrastructure. Pay only for the routing, not the tokens.
Start using the Gateway immediately with no barriers to entry. Use Kilo Gateway automatically inside IDE, CLI, and Cloud, or call it directly from your own apps via API.
Everything a production app needs from a model gateway: resilience, a standard integration surface, and visibility into routing decisions.
Built-in abuse detection and heuristic checks protect your implementation from anomalous traffic patterns.
Drop-in compatible with standard AI SDKs (like Vercel AI SDK). Switch providers without rewriting your codebase.
Our infrastructure handles the routing complexity, load balancing requests between direct providers and gateways like OpenRouter/Vercel to ensure maximum uptime.
Route each request to the best model available while keeping routing decisions, spend, and model behavior visible and controllable.
Auto Model
Kilo Gateway gives Auto Model access to the hosted catalog and routing surface. Kilo routes each task to the best model available while keeping the routing decision, spend, and model behavior visible and controllable.
Learn about Auto ModelTeam & governance
Enterprise provider configuration gives teams a concrete way to turn policy into day-to-day developer defaults. Admins can approve providers, restrict model access, and keep unsupported routes out of the engineering workflow.

Use Kilo Gateway to route requests across approved models and providers, with fallback paths and orchestration handled centrally.
Track model usage, latency, spend, BYOK activity, and adoption across teams from one place.
Approve providers, restrict model access, manage keys, and enforce identity controls so teams stay compliant without losing flexibility.
Flexible options to fit your needs, from a free tier to pay-as-you-go and Kilo Pass.
Free
Perfect for getting started
Pay-As-You-Go
For serious developers
Kilo Pass
Maximize your AI output per dollar.
Three steps stand between you and routing production traffic through Kilo Gateway.
Step 1
Generate your Kilo Gateway key or input your own provider keys.
Step 2
Point your Kilo apps or external applications to our Gateway endpoint.
Step 3
Gain insights into your model usage, latency, and costs through our unified dashboard.
Get your API key in seconds and start routing requests today.
We use this Gateway to power our own high-performance coding tools. The same infrastructure that handles millions of requests for Kilo Code is now available to you.