Envoy AI Gateway review
A Kubernetes-native open gateway for routing, governing, and observing LLM traffic.
Reviewed by iTechGuides Editors · Editorial team · Updated Oct 2026
Envoy AI Gateway is an open-source, Kubernetes-oriented gateway for applications that send traffic to generative AI services. It provides a unified API across hosted providers and self-hosted model endpoints, with provider authentication, upstream authorization, model and provider routing, automatic failover, rate limiting, budgets, and observability. Its strongest fit is for Kubernetes teams that want to control how LLM traffic moves across providers rather than integrate each provider separately.
The ecosystem coverage is broad. Integrations include OpenAI, AWS Bedrock, Azure OpenAI, Google Gemini, Google Vertex AI, Anthropic, Groq, Grok, Together AI, Cohere, Mistral, DeepInfra, DeepSeek, Hunyuan, and SambaNova. Routing can use weighted or prioritized backends, while traffic, usage, performance, and cost observability provide operational visibility. Support for self-hosted model endpoints also makes the gateway suitable for environments combining external APIs with models running under their own control.
Pricing is straightforward: Envoy AI Gateway is open source and has a free plan. The published product details describe capabilities rather than multiple paid tiers, so its value centers on the gateway software and the control it provides. Installation is Kubernetes-native through Helm charts, which aligns well with teams already operating Kubernetes but makes the product less suitable for organizations seeking a fully managed, non-Kubernetes gateway. Teams that need routing governance, fallback behavior, budget controls, and provider flexibility should consider it; teams wanting a hosted service with minimal platform ownership should choose an alternative. The project was renamed Agent Router on September 10, 2026.
Envoy AI Gateway pros and cons
- Where it wins
- Routes across a broad set of hosted and self-hosted model providers
- Adds failover, weighted routing, token limits, and usage budgets
- Helm-based Kubernetes installation with traffic and cost observability
- Where it doesn't
- Kubernetes-oriented deployment may not suit teams seeking a hosted gateway
- Self-hosted operation places deployment responsibility on the team
- Founded in 2024, making it a newer project
Envoy AI Gateway fact sheet, pricing and score →
Advertiser disclosure: iTechGuides is reader-supported. Vendors can pay for top positions in our rankings and for a place on other products' pages, and we may earn a commission when you click some links. Bought places stand at the number bought; this page and our disclosure say so. Paying vendors may get a fuller, more detailed review from our editors, and that review can change their score. How we rank.
Last updated · How we research and update