Policy Llm

Authorizes body-addressed inference requests against llm: policy routes

Versions marked “overview” do not contain this page. Selecting one opens that version’s documentation overview.

Category: Setup-dependent integration
Task: Authorizes body-addressed inference requests against llm: policy routes

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

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

Download the source file.

Companion resources from the same snapshot:

# Inference Authorization
#
# Authorizes body-addressed inference requests against `llm:` policy routes.
# The proxy reads the top-level `model`; missing, unlisted, and ambiguous
# models fail closed by default. No protocol classifier is required.
#
# Point `config_path` at an operator-supplied policy. The integration fixture
# at `tests/integration/fixtures/llm-policy.yaml` provides a local example.
#
# Requires building with the `policy-engine` feature, which is off
# by default:
#
# Usage:
#   cargo run -p praxis-proxy --features policy-engine -- -c examples/configs/security/policy-llm.yaml
#
listeners:
  - name: default
    address: "127.0.0.1:8080"
    filter_chains:
      - main

filter_chains:
  - name: main
    filters:
      - filter: policy
        config_path: /etc/praxis/llm-policy.yaml
        llm:
          require_model: true
          provider: openai

      # URL-addressed providers do not expose the model this filter evaluates.
      - filter: router
        routes:
          - path_prefix: "/v1/"
            cluster: inference

      - filter: load_balancer
        clusters:
          - name: inference
            endpoints:
              - "127.0.0.1:3000"

insecure_options:
  allow_private_endpoints: true