Model To Provider

Map a stable public model ID to an existing upstream cluster and target model

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

Category: Setup-dependent integration
Task: Map a stable public model ID to an existing upstream cluster and target model

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

Run it: Use ghcr.io/praxis-proxy/ai:0.8.0 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.

# Model-to-upstream resolution
#
# Map a stable public model ID to an existing upstream cluster and target model.
# Unmapped models and paths use the fallback backend without a body rewrite.
# Responses retain the target model unless a provider response adapter restores
# the public ID. Cluster application-provider metadata is optional for routing.
#
filter_chains:
  - name: mapped-model
    filters:
      - filter: model_to_provider
        fallback_upstream: fallback
        models:
          - model: claude-sonnet-4-5
            upstream: vertex
            target_model: claude-sonnet-4-5@20250929
            paths:
              - /v1/messages
              - /v1/messages/count_tokens
      - filter: load_balancer
        clusters:
          - name: vertex
            endpoints:
              - "127.0.0.1:9001"
          - name: fallback
            endpoints:
              - "127.0.0.1:9002"

insecure_options:
  allow_private_endpoints: true