API Reference¶

Package: pydantic_ai_web_models

This section documents all public exports of the package. Import them directly from the top-level module:

from pydantic_ai_web_models import (
    WebModel,
    TemporalConfig,
    get_default_config,
    set_default_config,
    AVAILABLE_MODELS,
    WebModelError,
    TemporalConnectionError,
    WorkflowExecutionError,
    JSONParseError,
)

Public Exports¶

Name	Kind	Description
`WebModel`	Class	The main model class. Implements the pydantic-ai `Model` interface and routes requests to Temporal.
`TemporalConfig`	Class	pydantic-settings `BaseSettings` subclass that holds all Temporal connection parameters.
`get_default_config`	Function	Returns the current module-level `TemporalConfig` instance.
`set_default_config`	Function	Replaces the module-level `TemporalConfig`. Takes effect for all subsequently created `WebModel` instances.
`AVAILABLE_MODELS`	Dict	Mapping of provider name to list of supported model name strings.
`WebModelError`	Exception	Base class for all exceptions raised by this library.
`TemporalConnectionError`	Exception	Raised when the library cannot connect to the Temporal server.
`WorkflowExecutionError`	Exception	Raised when the Temporal workflow returns an error. Carries `workflow_id`.
`JSONParseError`	Exception	Raised when structured-output JSON extraction fails. Carries `raw_text`.

Auto-Registration¶

Provider registration on import

Importing pydantic_ai_web_models automatically patches pydantic_ai.models.infer_model to recognise the "google-web:" and "openai-web:" provider prefixes. This means you can pass these model strings directly to Agent(model=...) without explicitly constructing a WebModel instance.

The registration happens once at import time and is idempotent — importing the module multiple times is safe.

import pydantic_ai_web_models  # registration happens here

from pydantic_ai import Agent
agent = Agent(model="google-web:gemini-3-flash")  # just works