Skip to main content

Adapters

OpenAI Adapter​

From docker-compose-base.yml:

adapter-openai:
  image: epam/ai-dial-adapter-openai:0.43.2
  platform: linux/amd64
  environment:
    DIAL_URL: "http://core:8080"
    WEB_CONCURRENCY: "3"
    DALLE3_DEPLOYMENTS: "dall-e-3"

Note

This sample runs the standalone DIAL OpenAI Adapter, which will be deprecated. For your own deployments, use DIAL Model Provider, which serves the same adapter under /compat/openai and reads the same variables. To move a configuration, see Migrate from the standalone adapters.

Environment variables​

VariableDescription
DIAL_URLDIAL Core internal URL. Used for file storage API calls.
WEB_CONCURRENCYNumber of Uvicorn worker processes.
DALLE3_DEPLOYMENTSComma-separated Azure OpenAI deployment names for DALL-E 3.
note

The 2.x stack no longer sets GPT4O_DEPLOYMENTS, GPT4O_MINI_DEPLOYMENTS, or DIAL_USE_FILE_STORAGE. Per-model deployment names are declared in config.json instead; DALLE3_DEPLOYMENTS remains because image models still need the adapter to know which deployments return images. See Adapter configuration for variables the adapter still accepts but this stack does not set.

See Adapter configuration for the full adapter environment variable reference.

Further reading​