Vllm Agentic Api

vLLM Agentic API: https://github.com/vllm-project/agentic-api

Category: Setup-dependent integration
Task: vLLM Agentic API: https://github.com/vllm-project/agentic-api

Prerequisites: The external service, credentials, or certificates referenced by this configuration.

Run it: Use ghcr.io/praxis-proxy/ai:0.4.1 and follow the container quickstart to mount and start the configuration.

This configuration comes from the selected release. The example has not been run here; external services are not bundled.

Download the source file.

# vLLM Agentic API Responses passthrough
#
# vLLM Agentic API: https://github.com/vllm-project/agentic-api
#
# Agentic API owns previous_response_id hydration, response persistence,
# web-search credentials, tool execution, and the model loop. Praxis only
# classifies and routes this traffic, preserving the original request body.
#
# This pipeline intentionally does not include openai_response_store,
# openai_responses_rehydrate, responses_proxy, openai_web_search,
# openai_tool_parse, or openai_stream_events.

listeners:
  - name: agentic-api-gateway
    address: "127.0.0.1:8080"
    filter_chains: [agentic-api-passthrough]

filter_chains:
  - name: agentic-api-passthrough
    filters:
      - filter: openai_responses_format
        on_invalid: continue

      - filter: router
        routes:
          - path: "/v1/responses"
            cluster: "vllm-agentic-api"

      - filter: load_balancer
        clusters:
          - name: "vllm-agentic-api"
            endpoints:
              - "127.0.0.1:3001"

insecure_options:
  allow_private_endpoints: true # example proxies to local backends