Kilo Gateway
OpenAI-compatible inference API that routes requests to 500+ models from 60+ providers at provider rates, with bring-your-own-key support, per-request cost tracking and organization model and provider controls.
Find alternatives to Kilo Gateway
What it does
- Model inference
Features 6
- Bring your own keyPlug in existing provider API keys and use the gateway as central infrastructure, paying only for routing.
- Standardized SDK integrationDrop-in compatible with standard AI SDKs such as the Vercel AI SDK.
- Auto Model routingRoutes each request to the best available model while keeping routing decisions and spend visible.
- Abuse detectionBuilt-in abuse detection and heuristic checks guard against anomalous traffic patterns.
- Usage insightsTracks model usage, latency, spend, BYOK activity and adoption across teams.
- Enterprise controlsAdmins approve providers, restrict model access, manage keys and enforce identity controls.
Sources
- Pricing
-
Gateway page read 2026-10-11: "Free $0 /forever Access to default free models BYOK support Basic analytics"; "Pay-As-You-Go ... All models at market rates No markup on BYOK"; "Kilo Pass ... From $19 /month A monthly AI token subscription With up to 50% bonus credits included". Pricing page: "Credit purchase processing fee 5%".
- Description
-
Official documentation · 11 Oct 2026
"The Kilo AI Gateway provides a unified, OpenAI-compatible API to access hundreds of AI models through a single endpoint"; "BYOK (Bring Your Own Key)"; "Per-request cost and token tracking"; "Model allow lists, provider restrictions, per-user daily spending limits". Pricing page: "500+ models across 60+ providers", "Pay exact provider rates with no markup".
- What it does
-
Official documentation · 11 Oct 2026
"Route LLM requests to any provider, anywhere."; load balancing between providers for uptime.
- Sold within
-
Sold on its own through its own API key and rate card: "Use Kilo Gateway to power your own apps", "Get API Key", tiers Free "$0 /forever", Pay-As-You-Go and Kilo Pass.
- Name
-
Official documentation · 11 Oct 2026
Heading: "Kilo Gateway The Universal Gateway for AI Inference".
- URL
-
Official documentation · 11 Oct 2026
Kilo's own product page, titled "Kilo Gateway - Universal AI Inference API".
- Deployment
-
Official documentation · 11 Oct 2026
Hosted endpoint at https://api.kilo.ai/api/gateway.
- Interfaces
-
Official documentation · 11 Oct 2026
"OpenAI-compatible API" at a single base URL; works with the Vercel AI SDK, the OpenAI SDK "or any OpenAI-compatible client".
- Intended for
-
Official documentation · 11 Oct 2026
Developers powering their own apps, and teams with "Enterprise provider configuration".
Also from Kilo Code
-
Kilo Code AI agent
AI coding agent for VS Code, JetBrains IDEs, CLI and cloud platforms, with modes for coding, architecture planning, debugging and code review, access to 500+ models and zero AI-inference markup (bring-your-own-key supported).
-
Kilo Cloud Agents AI agent
Runs the Kilo coding agent in an isolated cloud Linux container started from a browser or phone: it clones the repository, works on its own branch and commits and pushes after every message, and can be started by webhook or on a schedule.
-
Kilo Code Reviewer Developer tool
Reviews pull requests on GitHub or GitLab (Bitbucket coming soon) and posts inline comments on bugs, security issues, performance and style, with a strictness setting and team instructions, and can also review uncommitted changes locally in VS Code or JetBrains.
-
Security Agent AI agent
Turns GitHub Dependabot alerts into tracked findings, uses an AI triage model and a sandboxed analysis of the repository to decide whether each vulnerable dependency is actually reachable, opens remediation pull requests and tracks remediation deadlines with audit reports.
Open Kilo Code in the directory
Something wrong here? Send a correction — quote this product id: kilo-gateway.