Kong
API gateway platform extended with AI Gateway for governing LLM traffic.
About
Kong is a widely deployed API gateway and connectivity platform; its AI Gateway turns the same gateway into a control plane for LLM, MCP, and agent-to-agent traffic. Teams get a unified endpoint with multi-provider routing and failover, semantic caching, prompt guards, PII sanitization, token quotas, and AI-specific analytics, without changing application code.
Kong AI Gateway 2.0 reached general availability in September 2026 with a dedicated runtime and unified management for models, MCP servers, and agents, and the 3.14 release added an Agent Gateway for A2A traffic governance. It integrates with providers including OpenAI, Anthropic, AWS Bedrock, and Google Vertex AI, and is managed through Kong Konnect.
What it can do
- Unified routing across 6+ LLM providers with failover
- Semantic caching and prompt compression
- Prompt guards, PII sanitization, and token quotas
- MCP and agent-to-agent traffic governance
- Plugin ecosystem and AI-specific analytics
Setup
Deploy via Kong Konnect with the AI Gateway quickstart script, or self-host the gateway data plane with Docker.
Prerequisites
LLM provider API keys, and a Konnect account for managed control-plane features.
Use cases
Works with
Trust score — 8/10
Scored 8/10 (verified tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 108; Trending score 82/100 from researcher assessment.
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.