vLLM provider

You can proxy requests to vLLM AI models through AI Gateway by creating AI Model Provider and AI Model entities. This reference documents all supported AI capabilities, configuration requirements, and provider-specific details needed for proper integration.

AI Gateway automatically routes requests to the appropriate vLLM API endpoints. The following table shows the upstream paths used for each capability.

The following tables show the AI capabilities supported by the vLLM provider when configuring AI Models.

By default, AI Gateway uses the path templates shown in the tables below (e.g., /chat/completions, /embeddings, etc.). To customize these paths, configure the config.paths field in your AI Model entity. Custom paths take the form {configured_path}/{template_path} — for example, if you set a custom path of /v2, requests to /embeddings would be routed to /v2/embeddings.

Text generation

Support for vLLM text generation capabilities:

The base URL is $UPSTREAM_URL.

AI Gateway uses this URL automatically. You only need to configure a URL if you’re using a self-hosted or vLLM-compatible endpoint, in which case set the upstream_url option in your AI Model configuration.

To use vLLM with AI Gateway, configure a new AI Model Provider. You can then access supported AI Models from vLLM.

Here’s a minimal configuration for chat completions:

curl -X POST https://{region}.api.konghq.com/v1/ai-gateways/{AIGatewayId}/model-providers \
    --header "accept: application/json" \
    --header "Content-Type: application/json" \
    --header "Authorization: Bearer $KONNECT_TOKEN" \
    --data '
    {
      "display_name": "vllm Production",
      "name": "my-vllm-account",
      "type": "vllm",
      "config": {
        "auth": {
          "type": "basic"
        }
      }
    }
    '

ai_gateway_model_providers:
  - ref: my-vllm-account
    ai_gateway: !lookup {id: !env AI_GATEWAY_ID}
    display_name: vllm Production
    name: my-vllm-account
    type: vllm
    config:
      auth:
        type: basic

A target is an entry in the targets array on the AI Model entity, not the AI Model Provider. Beyond the common target options (name, provider, weight), a target routing to vLLM requires:

• upstream_url: The URL of your self-hosted vLLM server.

targets:
  - name: my-vllm-model
    provider: my-vllm-account
    config:
      type: vllm
      upstream_url: http://my-vllm-server.internal:8000