---
id: "helicone-gateway"
type: "component"
kind: "api-gateway"
provider: "default"
category: "edge"
docs: "https://docs.helicone.ai/gateway/provider-routing"
url: "https://localhost:3000/catalogue/components/helicone-gateway"
---

# Helicone AI Gateway

Open-source OpenAI-compatible router to many providers with automatic failover, caching and observability.

| | |
|---|---|
| Provider | [default](https://localhost:3000/catalogue?provider=default) |
| Kind | [API gateway](https://localhost:3000/catalogue/kinds/api-gateway) |
| Category | Edge |
| Official docs | [https://docs.helicone.ai/gateway/provider-routing](https://docs.helicone.ai/gateway/provider-routing) |

## Properties

| property | type | values | required | kind default | this component |
|---|---|---|---|---|---|
| `auth` | enum | `none`, `api-key`, `jwt`, `oauth2`, `mtls` | no | "none" | **"api-key"** |
| `cacheTtlSeconds` | integer | — | no | 604800 | inherited |
| `caching` | boolean | — | no | false | inherited |
| `concurrencyLimit` | boolean | — | no | false | inherited |
| `fallbacks` | boolean | — | no | true | inherited |
| `idempotencyKeys` (advanced) | boolean | — | no | false | inherited |
| `instances` | integer | — | no | — | inherited |
| `keyTtlHours` (advanced) | integer | — | no | — | inherited |
| `managed` | boolean | — | no | false | **true** |
| `protocol` | enum | `http`, `grpc`, `websocket`, `fix`, `ouch` | no | "http" | **"http"** |
| `rateLimiting` | boolean | — | no | false | **false** |
| `routingStrategy` | enum | `cheapest-first`, `provider-order` | no | "cheapest-first" | inherited |
| `tls` | boolean | — | no | true | **true** |

Bold values are the ones this component pins.

## Rules that apply · 1

### Synchronous call chain too deep `sync-chain-too-deep` (warning)

- Page: https://localhost:3000/catalogue/rules/sync-chain-too-deep.md
- Why: Every consecutive synchronous hop from this entry point spends part of the stated p95 latency budget, and the chain's failures multiply. More than three hops is a budget nobody is tracking.
- Applies to: every node of kind api-gateway
- Message: Entry point '{{node.id}}' starts a synchronous call chain deeper than 3 hops while the design states a p95 latency target.

It fires when:

- all of these hold:
  - the design requirement `latency.p95Ms` has a value
  - the longest chain of synchronous calls starting at the node is more than 3 hops long

## Other products of the same kind · 40

- [API Gateway](https://localhost:3000/catalogue/components/alibaba-api-gateway.md) `alibaba-api-gateway` — alibabacloud
- [Amazon API Gateway](https://localhost:3000/catalogue/components/aws-api-gateway.md) `aws-api-gateway` — aws
- [AWS AppSync GraphQL API](https://localhost:3000/catalogue/components/aws-appsync.md) `aws-appsync` — aws
- [Azure API Management](https://localhost:3000/catalogue/components/azure-api-management.md) `azure-api-management` — azure
- [AI Gateway](https://localhost:3000/catalogue/components/cloudflare-ai-gateway.md) `cloudflare-ai-gateway` — cloudflare
- [Cloudflare API Shield](https://localhost:3000/catalogue/components/cloudflare-api-shield.md) `cloudflare-api-shield` — cloudflare
- [MCP Server Portal](https://localhost:3000/catalogue/components/cloudflare-mcp-portal.md) `cloudflare-mcp-portal` — cloudflare
- [Cloudflare Privacy Gateway](https://localhost:3000/catalogue/components/cloudflare-privacy-gateway.md) `cloudflare-privacy-gateway` — cloudflare
- [Generic API gateway](https://localhost:3000/catalogue/components/api-gateway.md) `api-gateway` — default
- [Apache APISIX](https://localhost:3000/catalogue/components/apisix.md) `apisix` — default
- [Apollo GraphOS](https://localhost:3000/catalogue/components/apollo-graphos.md) `apollo-graphos` — default
- [Easegress](https://localhost:3000/catalogue/components/easegress.md) `easegress` — default
- [Emissary-ingress](https://localhost:3000/catalogue/components/emissary.md) `emissary` — default
- [Envoy Gateway](https://localhost:3000/catalogue/components/envoy-gateway.md) `envoy-gateway` — default
- [Gloo Gateway](https://localhost:3000/catalogue/components/gloo.md) `gloo` — default
- [Hasura](https://localhost:3000/catalogue/components/hasura.md) `hasura` — default
- [Higress](https://localhost:3000/catalogue/components/higress.md) `higress` — default
- [Kong Gateway](https://localhost:3000/catalogue/components/kong.md) `kong` — default
- [Kong AI Gateway](https://localhost:3000/catalogue/components/kong-ai-gateway.md) `kong-ai-gateway` — default
- [Kong Konnect](https://localhost:3000/catalogue/components/kong-konnect.md) `kong-konnect` — default
- [KrakenD](https://localhost:3000/catalogue/components/krakend.md) `krakend` — default
- [LiteLLM Proxy](https://localhost:3000/catalogue/components/litellm.md) `litellm` — default
- [Martian Gateway](https://localhost:3000/catalogue/components/martian.md) `martian` — default
- [Not Diamond](https://localhost:3000/catalogue/components/not-diamond.md) `not-diamond` — default
- [Portkey AI Gateway](https://localhost:3000/catalogue/components/portkey.md) `portkey` — default
- [Requesty](https://localhost:3000/catalogue/components/requesty.md) `requesty` — default
- [Spring Cloud Gateway](https://localhost:3000/catalogue/components/spring-cloud-gateway.md) `spring-cloud-gateway` — default
- [Tyk Gateway](https://localhost:3000/catalogue/components/tyk.md) `tyk` — default
- [Unify](https://localhost:3000/catalogue/components/unify-ai.md) `unify-ai` — default
- [Vercel AI Gateway](https://localhost:3000/catalogue/components/vercel-ai-gateway.md) `vercel-ai-gateway` — default
- [WSO2 API Manager](https://localhost:3000/catalogue/components/wso2-api-manager.md) `wso2-api-manager` — default
- [Zuul](https://localhost:3000/catalogue/components/zuul.md) `zuul` — default
- [Apigee](https://localhost:3000/catalogue/components/apigee.md) `apigee` — gcp
- [Agent Gateway](https://localhost:3000/catalogue/components/google-agent-gateway.md) `google-agent-gateway` — gcp
- [API Gateway](https://localhost:3000/catalogue/components/google-api-gateway.md) `google-api-gateway` — gcp
- [Apigee hybrid](https://localhost:3000/catalogue/components/google-apigee-hybrid.md) `google-apigee-hybrid` — gcp
- [Cloud Endpoints](https://localhost:3000/catalogue/components/google-cloud-endpoints.md) `google-cloud-endpoints` — gcp
- [Firebase Data Connect](https://localhost:3000/catalogue/components/google-firebase-data-connect.md) `google-firebase-data-connect` — gcp
- [IBM API Connect](https://localhost:3000/catalogue/components/ibm-api-connect.md) `ibm-api-connect` — ibm
- [OCI API Gateway](https://localhost:3000/catalogue/components/oci-api-gateway.md) `oci-api-gateway` — oci

## Use it in a design

```yaml
nodes:
  - { id: "helicone-gateway", component: "helicone-gateway" }
```

## Source (`component` YAML, verbatim)

```yaml
# yaml-language-server: $schema=../../schema/component.schema.json
# Helicone AI Gateway: OpenAI-compatible router to 100+ providers with observability, open source; in
# maintenance mode since Mintlify acquired Helicone in March 2026. Per "Provider Routing", it routes to
# the cheapest provider first and fails over automatically; per "LLM Caching", caching is off until the
# Helicone-Cache-Enabled header is sent (max-age 604800); rate limits need the Helicone-RateLimit-Policy header.
schemaVersion: "1"
type: "component"
id: "helicone-gateway"
kind: "api-gateway"
provider: "default"
docs: "https://docs.helicone.ai/gateway/provider-routing"
defaults:
  managed: true
  auth: "api-key"
  protocol: "http"
  tls: true
  rateLimiting: false
properties:
  routingStrategy:
    type: "enum"
    values: ["cheapest-first", "provider-order"]
    default: "cheapest-first"
  fallbacks:
    type: "boolean"
    default: true
  caching:
    type: "boolean"
    default: false
  cacheTtlSeconds:
    type: "integer"
    default: 604800
```

Human page: https://localhost:3000/catalogue/components/helicone-gateway