diff --git a/app/config.py b/app/config.py
index 0def557c..f9a18df5 100644
--- a/app/config.py
+++ b/app/config.py
@@ -15,6 +15,28 @@ class Settings(BaseSettings):
     openai_api_key: str
     openai_base_url: str = "https://api.openai.com/v1"  # Default to OpenAI's endpoint
     openai_model: str = "gpt-4o-mini"  # Default model
+
+    # AI provider abstraction layer
+    # Supported values: openai, azure, anthropic, gemini, ollama, openrouter, litellm
+    ai_provider: str = "openai"
+    # Override model for any provider; falls back to openai_model when not set
+    ai_model: Optional[str] = None
+
+    # Anthropic Claude settings (used when ai_provider="anthropic")
+    anthropic_api_key: Optional[str] = None
+
+    # Google Gemini settings (used when ai_provider="gemini")
+    gemini_api_key: Optional[str] = None
+
+    # Ollama local LLM settings (used when ai_provider="ollama")
+    ollama_base_url: str = "http://localhost:11434"
+
+    # OpenRouter settings (used when ai_provider="openrouter")
+    openrouter_api_key: Optional[str] = None
+    openrouter_base_url: str = "https://openrouter.ai/api/v1"
+
+    # Azure OpenAI API version (used when ai_provider="azure")
+    azure_openai_api_version: str = "2024-02-01"
     workdir: str
     debug: bool = False  # Default to False
 
diff --git a/app/tasks/extract_metadata_with_gpt.py b/app/tasks/extract_metadata_with_gpt.py
index c6fd33f9..162a75f3 100644
--- a/app/tasks/extract_metadata_with_gpt.py
+++ b/app/tasks/extract_metadata_with_gpt.py
@@ -5,8 +5,6 @@ import logging
 import os
 import re
 
-import openai
-
 # Import the shared Celery instance
 from app.celery_app import celery
 from app.config import settings
@@ -15,17 +13,10 @@ from app.models import FileRecord
 from app.tasks.embed_metadata_into_pdf import embed_metadata_into_pdf
 from app.tasks.retry_config import BaseTaskWithRetry
 from app.utils import log_task_progress
+from app.utils.ai_provider import get_ai_provider
 
 logger = logging.getLogger(__name__)
 
-# Initialize OpenAI client dynamically with better error handling
-try:
-    client = openai.OpenAI(api_key=settings.openai_api_key, base_url=settings.openai_base_url)
-    logger.info("OpenAI client initialized successfully")
-except Exception as e:
-    logger.error(f"Failed to initialize OpenAI client: {e}")
-    client = None
-
 
 def extract_json_from_text(text):
     """
@@ -114,23 +105,24 @@ def extract_metadata_with_gpt(self, filename: str, cleaned_text: str, file_id: i
 
     try:
         logger.info(f"[{task_id}] Sending classification request for {filename}...")
-        log_task_progress(task_id, "call_openai", "in_progress", "Calling OpenAI API", file_id=file_id)
-        completion = client.chat.completions.create(
-            model=settings.openai_model,
+        log_task_progress(task_id, "call_ai_provider", "in_progress", "Calling AI provider API", file_id=file_id)
+        provider = get_ai_provider()
+        model = settings.ai_model or settings.openai_model
+        content = provider.chat_completion(
             messages=[
                 {"role": "system", "content": "You are an intelligent document classifier."},
                 {"role": "user", "content": prompt},
             ],
+            model=model,
             temperature=0,
         )
 
-        content = completion.choices[0].message.content
         logger.info(f"[{task_id}] Raw classification response for {filename}: {content[:200]}...")
         log_task_progress(
             task_id,
-            "call_openai",
+            "call_ai_provider",
             "success",
-            "Received OpenAI response",
+            "Received AI provider response",
             file_id=file_id,
             detail=f"Raw classification response:\n{content}",
         )
@@ -187,13 +179,13 @@ def extract_metadata_with_gpt(self, filename: str, cleaned_text: str, file_id: i
         return {"s3_file": os.path.basename(filename), "metadata": metadata}
 
     except Exception as e:
-        logger.exception(f"[{task_id}] OpenAI classification failed for {filename}: {e}")
+        logger.exception(f"[{task_id}] AI provider classification failed for {filename}: {e}")
         log_task_progress(
             task_id,
             "extract_metadata_with_gpt",
             "failure",
             f"Exception: {str(e)}",
             file_id=file_id,
-            detail=f"OpenAI classification failed for {filename}.\nException: {str(e)}",
+            detail=f"AI provider classification failed for {filename}.\nException: {str(e)}",
         )
         return {}
diff --git a/app/tasks/refine_text_with_gpt.py b/app/tasks/refine_text_with_gpt.py
index d67f4c94..62b74c14 100644
--- a/app/tasks/refine_text_with_gpt.py
+++ b/app/tasks/refine_text_with_gpt.py
@@ -2,32 +2,29 @@
 
 import logging
 
-import openai
-
 # Import the shared Celery instance
 from app.celery_app import celery
 from app.config import settings
 from app.tasks.retry_config import BaseTaskWithRetry
 from app.utils import log_task_progress
+from app.utils.ai_provider import get_ai_provider
 
 logger = logging.getLogger(__name__)
 
-# Initialize OpenAI client dynamically
-client = openai.OpenAI(api_key=settings.openai_api_key, base_url=settings.openai_base_url)
-
 
 @celery.task(base=BaseTaskWithRetry, bind=True)
 def refine_text_with_gpt(self, filename: str, raw_text: str):
-    """Uses OpenAI to clean and refine OCR text."""
+    """Uses the configured AI provider to clean and refine OCR text."""
     task_id = self.request.id
     logger.info(f"[{task_id}] Starting OCR text refinement for: {filename}")
     log_task_progress(task_id, "refine_text_with_gpt", "in_progress", f"Refining OCR text for {filename}")
 
     try:
-        log_task_progress(task_id, "call_openai", "in_progress", "Calling OpenAI for text refinement")
+        log_task_progress(task_id, "call_ai_provider", "in_progress", "Calling AI provider for text refinement")
 
-        response = client.chat.completions.create(
-            model=settings.openai_model,
+        provider = get_ai_provider()
+        model = settings.ai_model or settings.openai_model
+        cleaned_text = provider.chat_completion(
             messages=[
                 {
                     "role": "system",
@@ -38,16 +35,15 @@ def refine_text_with_gpt(self, filename: str, raw_text: str):
                 },
                 {"role": "user", "content": raw_text},
             ],
+            model=model,
         )
 
-        cleaned_text = response.choices[0].message.content
-
         logger.info(f"[{task_id}] Text refinement complete for {filename}: {len(cleaned_text)} characters")
         log_task_progress(
             task_id,
-            "call_openai",
+            "call_ai_provider",
             "success",
-            "Received refined text from OpenAI",
+            "Received refined text from AI provider",
             detail=f"Input: {len(raw_text)} chars → Output: {len(cleaned_text)} chars",
         )
 
diff --git a/app/utils/ai_provider.py b/app/utils/ai_provider.py
new file mode 100644
index 00000000..a98cae90
--- /dev/null
+++ b/app/utils/ai_provider.py
@@ -0,0 +1,345 @@
+#!/usr/bin/env python3
+"""AI provider abstraction layer for DocuElevate.
+
+This module provides a pluggable abstraction for various AI model providers,
+allowing the platform to work with OpenAI, Azure OpenAI, Anthropic Claude,
+Google Gemini, Ollama (local LLMs), OpenRouter, and any LiteLLM-compatible
+provider without being locked to a single vendor.
+
+Provider selection is controlled by the ``AI_PROVIDER`` environment variable.
+See the Configuration Guide for full details on each provider's settings.
+"""
+
+import logging
+from abc import ABC, abstractmethod
+from typing import Any, Dict, List, Optional
+
+logger = logging.getLogger(__name__)
+
+
+class AIProvider(ABC):
+    """Abstract base class for AI chat completion providers.
+
+    All concrete providers must implement :meth:`chat_completion`, which
+    accepts a list of chat messages and returns the model's response as a
+    plain string.  The interface intentionally mirrors the OpenAI Chat
+    Completions API so that callers need no provider-specific knowledge.
+    """
+
+    @abstractmethod
+    def chat_completion(
+        self,
+        messages: List[Dict[str, str]],
+        model: str,
+        temperature: float = 0,
+        **kwargs: Any,
+    ) -> str:
+        """Get a chat completion from the AI provider.
+
+        Args:
+            messages: List of message dicts with ``role`` and ``content`` keys.
+            model: Model name/identifier to use (provider-specific format).
+            temperature: Sampling temperature (0–1). Default: 0 (deterministic).
+            **kwargs: Additional provider-specific arguments passed through.
+
+        Returns:
+            The model's response as a plain string.
+
+        Raises:
+            Exception: If the underlying API call fails.
+        """
+
+
+class OpenAIProvider(AIProvider):
+    """OpenAI provider using the ``openai`` Python SDK.
+
+    Also works as a drop-in for any OpenAI-compatible API endpoint, including
+    LocalAI and LM Studio.  Ollama and OpenRouter have dedicated providers with
+    sensible defaults, but this provider works for them too when a custom
+    ``base_url`` is supplied.
+    """
+
+    def __init__(self, api_key: str, base_url: Optional[str] = None) -> None:
+        import openai
+
+        self._client = openai.OpenAI(
+            api_key=api_key,
+            base_url=base_url or "https://api.openai.com/v1",
+        )
+
+    def chat_completion(
+        self,
+        messages: List[Dict[str, str]],
+        model: str,
+        temperature: float = 0,
+        **kwargs: Any,
+    ) -> str:
+        completion = self._client.chat.completions.create(
+            model=model,
+            messages=messages,
+            temperature=temperature,
+            **kwargs,
+        )
+        return completion.choices[0].message.content
+
+
+class AzureOpenAIProvider(AIProvider):
+    """Azure OpenAI provider using the ``openai`` Python SDK's Azure client."""
+
+    def __init__(self, api_key: str, azure_endpoint: str, api_version: str = "2024-02-01") -> None:
+        import openai
+
+        self._client = openai.AzureOpenAI(
+            api_key=api_key,
+            azure_endpoint=azure_endpoint,
+            api_version=api_version,
+        )
+
+    def chat_completion(
+        self,
+        messages: List[Dict[str, str]],
+        model: str,
+        temperature: float = 0,
+        **kwargs: Any,
+    ) -> str:
+        completion = self._client.chat.completions.create(
+            model=model,
+            messages=messages,
+            temperature=temperature,
+            **kwargs,
+        )
+        return completion.choices[0].message.content
+
+
+class AnthropicProvider(AIProvider):
+    """Anthropic Claude provider routed via LiteLLM.
+
+    Requires ``litellm`` to be installed.  Model names should be in Anthropic
+    format (e.g. ``claude-3-5-sonnet-20241022``); the ``anthropic/`` prefix is
+    added automatically when absent.
+    """
+
+    def __init__(self, api_key: str) -> None:
+        self._api_key = api_key
+
+    def chat_completion(
+        self,
+        messages: List[Dict[str, str]],
+        model: str,
+        temperature: float = 0,
+        **kwargs: Any,
+    ) -> str:
+        import litellm
+
+        model_name = model if model.startswith("anthropic/") else f"anthropic/{model}"
+        response = litellm.completion(
+            model=model_name,
+            messages=messages,
+            temperature=temperature,
+            api_key=self._api_key,
+            **kwargs,
+        )
+        return response.choices[0].message.content
+
+
+class GeminiProvider(AIProvider):
+    """Google Gemini provider routed via LiteLLM.
+
+    Requires ``litellm`` to be installed.  Model names should be in Gemini
+    format (e.g. ``gemini-1.5-pro``); the ``gemini/`` prefix is added
+    automatically when absent.
+    """
+
+    def __init__(self, api_key: str) -> None:
+        self._api_key = api_key
+
+    def chat_completion(
+        self,
+        messages: List[Dict[str, str]],
+        model: str,
+        temperature: float = 0,
+        **kwargs: Any,
+    ) -> str:
+        import litellm
+
+        model_name = model if model.startswith("gemini/") else f"gemini/{model}"
+        response = litellm.completion(
+            model=model_name,
+            messages=messages,
+            temperature=temperature,
+            api_key=self._api_key,
+            **kwargs,
+        )
+        return response.choices[0].message.content
+
+
+class OllamaProvider(AIProvider):
+    """Ollama local LLM provider via its OpenAI-compatible REST API.
+
+    Ollama exposes an OpenAI-compatible endpoint at ``/v1``.  Any model
+    pulled into your Ollama instance (e.g. ``llama3.2``, ``qwen2.5``,
+    ``phi3``) can be used directly by name.
+
+    For CPU-only deployments the recommended models are:
+
+    * ``llama3.2`` (3B) – good balance of speed and quality
+    * ``qwen2.5`` (3B/7B) – excellent at structured JSON output
+    * ``phi3`` (3.8B) – strong reasoning, fast on CPU
+
+    See https://ollama.com for installation and model management.
+    """
+
+    def __init__(self, base_url: str = "http://localhost:11434") -> None:
+        import openai
+
+        self._client = openai.OpenAI(
+            api_key="ollama",  # Ollama does not require a real API key
+            base_url=f"{base_url.rstrip('/')}/v1",
+        )
+
+    def chat_completion(
+        self,
+        messages: List[Dict[str, str]],
+        model: str,
+        temperature: float = 0,
+        **kwargs: Any,
+    ) -> str:
+        completion = self._client.chat.completions.create(
+            model=model,
+            messages=messages,
+            temperature=temperature,
+            **kwargs,
+        )
+        return completion.choices[0].message.content
+
+
+class OpenRouterProvider(AIProvider):
+    """OpenRouter AI aggregator (https://openrouter.ai).
+
+    OpenRouter provides access to 100+ models from OpenAI, Anthropic, Google,
+    Meta, Mistral, and many others through a single OpenAI-compatible endpoint.
+    Model names use the ``provider/model`` format (e.g.
+    ``anthropic/claude-3.5-sonnet``, ``google/gemini-pro``).
+    """
+
+    def __init__(self, api_key: str, base_url: str = "https://openrouter.ai/api/v1") -> None:
+        import openai
+
+        self._client = openai.OpenAI(
+            api_key=api_key,
+            base_url=base_url,
+        )
+
+    def chat_completion(
+        self,
+        messages: List[Dict[str, str]],
+        model: str,
+        temperature: float = 0,
+        **kwargs: Any,
+    ) -> str:
+        completion = self._client.chat.completions.create(
+            model=model,
+            messages=messages,
+            temperature=temperature,
+            **kwargs,
+        )
+        return completion.choices[0].message.content
+
+
+class LiteLLMProvider(AIProvider):
+    """LiteLLM provider – unified interface for 100+ LLMs.
+
+    LiteLLM (https://litellm.ai) translates calls to a single interface that
+    supports OpenAI, Azure, Anthropic, Gemini, Cohere, Ollama, and many more.
+    Use the LiteLLM model-string format ``provider/model`` (e.g.
+    ``openai/gpt-4o``, ``anthropic/claude-3-5-sonnet-20241022``,
+    ``ollama/llama3.2``).
+
+    This provider is useful when you want LiteLLM to handle all routing and
+    need features like automatic retries, fallbacks, or cost tracking.
+    """
+
+    def __init__(self, api_key: Optional[str] = None, api_base: Optional[str] = None) -> None:
+        self._api_key = api_key
+        self._api_base = api_base
+
+    def chat_completion(
+        self,
+        messages: List[Dict[str, str]],
+        model: str,
+        temperature: float = 0,
+        **kwargs: Any,
+    ) -> str:
+        import litellm
+
+        completion_kwargs: Dict[str, Any] = {
+            "model": model,
+            "messages": messages,
+            "temperature": temperature,
+        }
+        if self._api_key:
+            completion_kwargs["api_key"] = self._api_key
+        if self._api_base:
+            completion_kwargs["api_base"] = self._api_base
+        completion_kwargs.update(kwargs)
+        response = litellm.completion(**completion_kwargs)
+        return response.choices[0].message.content
+
+
+def get_ai_provider() -> AIProvider:
+    """Factory function that creates and returns the configured AI provider.
+
+    Reads ``settings.ai_provider`` (set via the ``AI_PROVIDER`` environment
+    variable) to select the provider implementation.  Provider-specific
+    credentials and URLs are read from their corresponding settings fields.
+
+    Returns:
+        An :class:`AIProvider` instance ready to serve chat completions.
+
+    Raises:
+        ValueError: If the configured provider name is not recognised.
+        ValueError: If required credentials for the selected provider are absent.
+    """
+    from app.config import settings
+
+    provider = settings.ai_provider.lower()
+    logger.debug(f"Creating AI provider: {provider}")
+
+    if provider == "openai":
+        return OpenAIProvider(
+            api_key=settings.openai_api_key,
+            base_url=settings.openai_base_url,
+        )
+    elif provider == "azure":
+        return AzureOpenAIProvider(
+            api_key=settings.openai_api_key,
+            azure_endpoint=settings.openai_base_url,
+            api_version=settings.azure_openai_api_version,
+        )
+    elif provider == "anthropic":
+        if not settings.anthropic_api_key:
+            raise ValueError("ANTHROPIC_API_KEY must be set when AI_PROVIDER='anthropic'")
+        return AnthropicProvider(api_key=settings.anthropic_api_key)
+    elif provider == "gemini":
+        if not settings.gemini_api_key:
+            raise ValueError("GEMINI_API_KEY must be set when AI_PROVIDER='gemini'")
+        return GeminiProvider(api_key=settings.gemini_api_key)
+    elif provider == "ollama":
+        return OllamaProvider(base_url=settings.ollama_base_url)
+    elif provider == "openrouter":
+        if not settings.openrouter_api_key:
+            raise ValueError("OPENROUTER_API_KEY must be set when AI_PROVIDER='openrouter'")
+        return OpenRouterProvider(
+            api_key=settings.openrouter_api_key,
+            base_url=settings.openrouter_base_url,
+        )
+    elif provider == "litellm":
+        return LiteLLMProvider(
+            api_key=settings.openai_api_key or None,
+            api_base=settings.openai_base_url if settings.openai_base_url != "https://api.openai.com/v1" else None,
+        )
+    else:
+        raise ValueError(
+            f"Unknown AI provider: '{provider}'. "
+            "Supported providers: openai, azure, anthropic, gemini, ollama, openrouter, litellm"
+        )
diff --git a/requirements.txt b/requirements.txt
index b1389fd6..a46fff38 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -34,4 +34,7 @@ boto3>=1.28.0
 paramiko>=3.4.0  # SSH/SFTP implementation for Python (LGPL license)
 
 # Notification service
-apprise>=1.4.0
+apprise>=1.4.0
+
+# AI provider aggregator - enables Anthropic, Gemini, Ollama, and 100+ LLM providers
+litellm>=1.0.0,<2.0.0
diff --git a/tests/test_ai_provider.py b/tests/test_ai_provider.py
new file mode 100644
index 00000000..e77caefa
--- /dev/null
+++ b/tests/test_ai_provider.py
@@ -0,0 +1,591 @@
+"""Unit tests for the AI provider abstraction layer (app/utils/ai_provider.py)."""
+
+from types import SimpleNamespace
+from unittest.mock import MagicMock, patch
+
+import pytest
+
+from app.utils.ai_provider import (
+    AIProvider,
+    AnthropicProvider,
+    AzureOpenAIProvider,
+    GeminiProvider,
+    LiteLLMProvider,
+    OllamaProvider,
+    OpenAIProvider,
+    OpenRouterProvider,
+    get_ai_provider,
+)
+
+
+# ---------------------------------------------------------------------------
+# Helpers
+# ---------------------------------------------------------------------------
+
+def _make_openai_response(content: str) -> MagicMock:
+    """Build a minimal mock that looks like an openai ChatCompletion response."""
+    msg = SimpleNamespace(content=content)
+    choice = SimpleNamespace(message=msg)
+    resp = MagicMock()
+    resp.choices = [choice]
+    return resp
+
+
+def _make_litellm_response(content: str) -> MagicMock:
+    """Build a minimal mock that looks like a litellm completion response."""
+    msg = SimpleNamespace(content=content)
+    choice = SimpleNamespace(message=msg)
+    resp = MagicMock()
+    resp.choices = [choice]
+    return resp
+
+
+# ---------------------------------------------------------------------------
+# Abstract base class
+# ---------------------------------------------------------------------------
+
+@pytest.mark.unit
+class TestAIProviderABC:
+    """Verify that AIProvider cannot be instantiated directly."""
+
+    def test_cannot_instantiate_directly(self):
+        """AIProvider is abstract and must be subclassed."""
+        with pytest.raises(TypeError):
+            AIProvider()  # type: ignore[abstract]
+
+    def test_subclass_must_implement_chat_completion(self):
+        """Concrete subclass without chat_completion raises TypeError."""
+
+        class Incomplete(AIProvider):
+            pass
+
+        with pytest.raises(TypeError):
+            Incomplete()  # type: ignore[abstract]
+
+
+# ---------------------------------------------------------------------------
+# OpenAIProvider
+# ---------------------------------------------------------------------------
+
+@pytest.mark.unit
+class TestOpenAIProvider:
+    """Tests for OpenAIProvider."""
+
+    @patch("openai.OpenAI")
+    def test_initialises_with_default_base_url(self, mock_openai_cls):
+        """Provider uses the OpenAI default endpoint when no base_url is given."""
+        OpenAIProvider(api_key="sk-test")
+        mock_openai_cls.assert_called_once_with(
+            api_key="sk-test",
+            base_url="https://api.openai.com/v1",
+        )
+
+    @patch("openai.OpenAI")
+    def test_initialises_with_custom_base_url(self, mock_openai_cls):
+        """Provider passes through a custom base_url."""
+        OpenAIProvider(api_key="sk-test", base_url="http://localhost:8000/v1")
+        mock_openai_cls.assert_called_once_with(
+            api_key="sk-test",
+            base_url="http://localhost:8000/v1",
+        )
+
+    @patch("openai.OpenAI")
+    def test_chat_completion_returns_content(self, mock_openai_cls):
+        """chat_completion extracts and returns the message content string."""
+        mock_client = MagicMock()
+        mock_client.chat.completions.create.return_value = _make_openai_response("Hello world")
+        mock_openai_cls.return_value = mock_client
+
+        provider = OpenAIProvider(api_key="sk-test")
+        result = provider.chat_completion(
+            messages=[{"role": "user", "content": "Hi"}],
+            model="gpt-4o-mini",
+            temperature=0,
+        )
+
+        assert result == "Hello world"
+        mock_client.chat.completions.create.assert_called_once_with(
+            model="gpt-4o-mini",
+            messages=[{"role": "user", "content": "Hi"}],
+            temperature=0,
+        )
+
+    @patch("openai.OpenAI")
+    def test_chat_completion_passes_kwargs(self, mock_openai_cls):
+        """Extra kwargs are forwarded to the underlying client."""
+        mock_client = MagicMock()
+        mock_client.chat.completions.create.return_value = _make_openai_response("ok")
+        mock_openai_cls.return_value = mock_client
+
+        provider = OpenAIProvider(api_key="sk-test")
+        provider.chat_completion(
+            messages=[{"role": "user", "content": "test"}],
+            model="gpt-4o",
+            temperature=0.7,
+            max_tokens=100,
+        )
+
+        call_kwargs = mock_client.chat.completions.create.call_args[1]
+        assert call_kwargs["max_tokens"] == 100
+        assert call_kwargs["temperature"] == 0.7
+
+
+# ---------------------------------------------------------------------------
+# AzureOpenAIProvider
+# ---------------------------------------------------------------------------
+
+@pytest.mark.unit
+class TestAzureOpenAIProvider:
+    """Tests for AzureOpenAIProvider."""
+
+    @patch("openai.AzureOpenAI")
+    def test_initialises_correctly(self, mock_azure_cls):
+        """Azure provider initialises with correct parameters."""
+        AzureOpenAIProvider(
+            api_key="azure-key",
+            azure_endpoint="https://my-resource.openai.azure.com",
+            api_version="2024-02-01",
+        )
+        mock_azure_cls.assert_called_once_with(
+            api_key="azure-key",
+            azure_endpoint="https://my-resource.openai.azure.com",
+            api_version="2024-02-01",
+        )
+
+    @patch("openai.AzureOpenAI")
+    def test_chat_completion_returns_content(self, mock_azure_cls):
+        """chat_completion returns the message content string."""
+        mock_client = MagicMock()
+        mock_client.chat.completions.create.return_value = _make_openai_response("Azure response")
+        mock_azure_cls.return_value = mock_client
+
+        provider = AzureOpenAIProvider(
+            api_key="key",
+            azure_endpoint="https://endpoint.openai.azure.com",
+        )
+        result = provider.chat_completion(
+            messages=[{"role": "user", "content": "test"}],
+            model="gpt-4",
+        )
+
+        assert result == "Azure response"
+
+
+# ---------------------------------------------------------------------------
+# AnthropicProvider
+# ---------------------------------------------------------------------------
+
+@pytest.mark.unit
+class TestAnthropicProvider:
+    """Tests for AnthropicProvider."""
+
+    @patch("litellm.completion")
+    def test_chat_completion_adds_anthropic_prefix(self, mock_completion):
+        """Provider prepends 'anthropic/' to bare model names."""
+        mock_completion.return_value = _make_litellm_response("Claude says hi")
+
+        provider = AnthropicProvider(api_key="ant-key")
+        result = provider.chat_completion(
+            messages=[{"role": "user", "content": "Hello"}],
+            model="claude-3-5-sonnet-20241022",
+        )
+
+        assert result == "Claude says hi"
+        call_kwargs = mock_completion.call_args[1]
+        assert call_kwargs["model"] == "anthropic/claude-3-5-sonnet-20241022"
+
+    @patch("litellm.completion")
+    def test_chat_completion_keeps_existing_prefix(self, mock_completion):
+        """Provider does not double-add 'anthropic/' if already present."""
+        mock_completion.return_value = _make_litellm_response("ok")
+
+        provider = AnthropicProvider(api_key="ant-key")
+        provider.chat_completion(
+            messages=[{"role": "user", "content": "test"}],
+            model="anthropic/claude-3-opus-20240229",
+        )
+
+        call_kwargs = mock_completion.call_args[1]
+        assert call_kwargs["model"] == "anthropic/claude-3-opus-20240229"
+
+    @patch("litellm.completion")
+    def test_chat_completion_passes_api_key(self, mock_completion):
+        """Provider passes the API key to litellm."""
+        mock_completion.return_value = _make_litellm_response("ok")
+
+        provider = AnthropicProvider(api_key="my-ant-key")
+        provider.chat_completion(
+            messages=[{"role": "user", "content": "test"}],
+            model="claude-3-5-sonnet-20241022",
+        )
+
+        call_kwargs = mock_completion.call_args[1]
+        assert call_kwargs["api_key"] == "my-ant-key"
+
+
+# ---------------------------------------------------------------------------
+# GeminiProvider
+# ---------------------------------------------------------------------------
+
+@pytest.mark.unit
+class TestGeminiProvider:
+    """Tests for GeminiProvider."""
+
+    @patch("litellm.completion")
+    def test_chat_completion_adds_gemini_prefix(self, mock_completion):
+        """Provider prepends 'gemini/' to bare model names."""
+        mock_completion.return_value = _make_litellm_response("Gemini says hi")
+
+        provider = GeminiProvider(api_key="gemini-key")
+        result = provider.chat_completion(
+            messages=[{"role": "user", "content": "Hello"}],
+            model="gemini-1.5-pro",
+        )
+
+        assert result == "Gemini says hi"
+        call_kwargs = mock_completion.call_args[1]
+        assert call_kwargs["model"] == "gemini/gemini-1.5-pro"
+
+    @patch("litellm.completion")
+    def test_chat_completion_keeps_existing_prefix(self, mock_completion):
+        """Provider does not double-add 'gemini/' if already present."""
+        mock_completion.return_value = _make_litellm_response("ok")
+
+        provider = GeminiProvider(api_key="gemini-key")
+        provider.chat_completion(
+            messages=[{"role": "user", "content": "test"}],
+            model="gemini/gemini-pro",
+        )
+
+        call_kwargs = mock_completion.call_args[1]
+        assert call_kwargs["model"] == "gemini/gemini-pro"
+
+
+# ---------------------------------------------------------------------------
+# OllamaProvider
+# ---------------------------------------------------------------------------
+
+@pytest.mark.unit
+class TestOllamaProvider:
+    """Tests for OllamaProvider."""
+
+    @patch("openai.OpenAI")
+    def test_initialises_with_default_base_url(self, mock_openai_cls):
+        """Provider appends /v1 to the default Ollama base URL."""
+        OllamaProvider()
+        mock_openai_cls.assert_called_once_with(
+            api_key="ollama",
+            base_url="http://localhost:11434/v1",
+        )
+
+    @patch("openai.OpenAI")
+    def test_initialises_with_custom_base_url(self, mock_openai_cls):
+        """Provider appends /v1 to a custom Ollama base URL."""
+        OllamaProvider(base_url="http://my-ollama:11434")
+        mock_openai_cls.assert_called_once_with(
+            api_key="ollama",
+            base_url="http://my-ollama:11434/v1",
+        )
+
+    @patch("openai.OpenAI")
+    def test_strips_trailing_slash_before_appending_v1(self, mock_openai_cls):
+        """Provider normalises trailing slashes before appending /v1."""
+        OllamaProvider(base_url="http://ollama:11434/")
+        call_kwargs = mock_openai_cls.call_args[1]
+        assert call_kwargs["base_url"] == "http://ollama:11434/v1"
+
+    @patch("openai.OpenAI")
+    def test_chat_completion_returns_content(self, mock_openai_cls):
+        """chat_completion returns the message content string."""
+        mock_client = MagicMock()
+        mock_client.chat.completions.create.return_value = _make_openai_response("Llama response")
+        mock_openai_cls.return_value = mock_client
+
+        provider = OllamaProvider()
+        result = provider.chat_completion(
+            messages=[{"role": "user", "content": "hello"}],
+            model="llama3.2",
+        )
+
+        assert result == "Llama response"
+
+
+# ---------------------------------------------------------------------------
+# OpenRouterProvider
+# ---------------------------------------------------------------------------
+
+@pytest.mark.unit
+class TestOpenRouterProvider:
+    """Tests for OpenRouterProvider."""
+
+    @patch("openai.OpenAI")
+    def test_initialises_with_default_base_url(self, mock_openai_cls):
+        """Provider uses the default OpenRouter endpoint."""
+        OpenRouterProvider(api_key="or-key")
+        mock_openai_cls.assert_called_once_with(
+            api_key="or-key",
+            base_url="https://openrouter.ai/api/v1",
+        )
+
+    @patch("openai.OpenAI")
+    def test_chat_completion_returns_content(self, mock_openai_cls):
+        """chat_completion returns the message content string."""
+        mock_client = MagicMock()
+        mock_client.chat.completions.create.return_value = _make_openai_response("Router response")
+        mock_openai_cls.return_value = mock_client
+
+        provider = OpenRouterProvider(api_key="or-key")
+        result = provider.chat_completion(
+            messages=[{"role": "user", "content": "hi"}],
+            model="anthropic/claude-3.5-sonnet",
+        )
+
+        assert result == "Router response"
+
+
+# ---------------------------------------------------------------------------
+# LiteLLMProvider
+# ---------------------------------------------------------------------------
+
+@pytest.mark.unit
+class TestLiteLLMProvider:
+    """Tests for LiteLLMProvider."""
+
+    @patch("litellm.completion")
+    def test_chat_completion_basic(self, mock_completion):
+        """Provider calls litellm.completion with correct arguments."""
+        mock_completion.return_value = _make_litellm_response("LiteLLM response")
+
+        provider = LiteLLMProvider()
+        result = provider.chat_completion(
+            messages=[{"role": "user", "content": "hello"}],
+            model="openai/gpt-4o",
+        )
+
+        assert result == "LiteLLM response"
+        call_kwargs = mock_completion.call_args[1]
+        assert call_kwargs["model"] == "openai/gpt-4o"
+        assert call_kwargs["temperature"] == 0
+
+    @patch("litellm.completion")
+    def test_chat_completion_with_api_key(self, mock_completion):
+        """Provider forwards api_key when set."""
+        mock_completion.return_value = _make_litellm_response("ok")
+
+        provider = LiteLLMProvider(api_key="my-key")
+        provider.chat_completion(
+            messages=[{"role": "user", "content": "test"}],
+            model="openai/gpt-4o",
+        )
+
+        call_kwargs = mock_completion.call_args[1]
+        assert call_kwargs["api_key"] == "my-key"
+
+    @patch("litellm.completion")
+    def test_chat_completion_with_api_base(self, mock_completion):
+        """Provider forwards api_base when set."""
+        mock_completion.return_value = _make_litellm_response("ok")
+
+        provider = LiteLLMProvider(api_base="http://my-proxy/v1")
+        provider.chat_completion(
+            messages=[{"role": "user", "content": "test"}],
+            model="openai/gpt-4o",
+        )
+
+        call_kwargs = mock_completion.call_args[1]
+        assert call_kwargs["api_base"] == "http://my-proxy/v1"
+
+    @patch("litellm.completion")
+    def test_chat_completion_omits_none_api_key(self, mock_completion):
+        """Provider does not include api_key when it is None."""
+        mock_completion.return_value = _make_litellm_response("ok")
+
+        provider = LiteLLMProvider()  # api_key=None
+        provider.chat_completion(
+            messages=[{"role": "user", "content": "test"}],
+            model="openai/gpt-4o",
+        )
+
+        call_kwargs = mock_completion.call_args[1]
+        assert "api_key" not in call_kwargs
+
+    @patch("litellm.completion")
+    def test_chat_completion_omits_none_api_base(self, mock_completion):
+        """Provider does not include api_base when it is None."""
+        mock_completion.return_value = _make_litellm_response("ok")
+
+        provider = LiteLLMProvider()  # api_base=None
+        provider.chat_completion(
+            messages=[{"role": "user", "content": "test"}],
+            model="openai/gpt-4o",
+        )
+
+        call_kwargs = mock_completion.call_args[1]
+        assert "api_base" not in call_kwargs
+
+
+# ---------------------------------------------------------------------------
+# get_ai_provider factory function
+# ---------------------------------------------------------------------------
+
+@pytest.mark.unit
+class TestGetAIProvider:
+    """Tests for the get_ai_provider factory function."""
+
+    def _mock_settings(self, **kwargs):
+        """Return a mock settings object with sensible defaults."""
+        defaults = {
+            "ai_provider": "openai",
+            "openai_api_key": "sk-test",
+            "openai_base_url": "https://api.openai.com/v1",
+            "azure_openai_api_version": "2024-02-01",
+            "anthropic_api_key": None,
+            "gemini_api_key": None,
+            "ollama_base_url": "http://localhost:11434",
+            "openrouter_api_key": None,
+            "openrouter_base_url": "https://openrouter.ai/api/v1",
+        }
+        defaults.update(kwargs)
+        mock_settings = MagicMock()
+        for key, value in defaults.items():
+            setattr(mock_settings, key, value)
+        return mock_settings
+
+    @patch("openai.OpenAI")
+    def test_returns_openai_provider_by_default(self, mock_openai_cls):
+        """get_ai_provider returns an OpenAIProvider for ai_provider='openai'."""
+        with patch("app.utils.ai_provider.settings", self._mock_settings(ai_provider="openai")):
+            provider = get_ai_provider()
+        assert isinstance(provider, OpenAIProvider)
+
+    @patch("openai.AzureOpenAI")
+    def test_returns_azure_provider(self, mock_azure_cls):
+        """get_ai_provider returns AzureOpenAIProvider for ai_provider='azure'."""
+        with patch(
+            "app.utils.ai_provider.settings",
+            self._mock_settings(
+                ai_provider="azure",
+                openai_base_url="https://my-resource.openai.azure.com",
+            ),
+        ):
+            provider = get_ai_provider()
+        assert isinstance(provider, AzureOpenAIProvider)
+
+    def test_returns_anthropic_provider(self):
+        """get_ai_provider returns AnthropicProvider for ai_provider='anthropic'."""
+        with patch(
+            "app.utils.ai_provider.settings",
+            self._mock_settings(ai_provider="anthropic", anthropic_api_key="ant-key"),
+        ):
+            provider = get_ai_provider()
+        assert isinstance(provider, AnthropicProvider)
+
+    def test_anthropic_raises_without_api_key(self):
+        """get_ai_provider raises ValueError when anthropic_api_key is missing."""
+        with patch(
+            "app.utils.ai_provider.settings",
+            self._mock_settings(ai_provider="anthropic", anthropic_api_key=None),
+        ):
+            with pytest.raises(ValueError, match="ANTHROPIC_API_KEY"):
+                get_ai_provider()
+
+    def test_returns_gemini_provider(self):
+        """get_ai_provider returns GeminiProvider for ai_provider='gemini'."""
+        with patch(
+            "app.utils.ai_provider.settings",
+            self._mock_settings(ai_provider="gemini", gemini_api_key="gemini-key"),
+        ):
+            provider = get_ai_provider()
+        assert isinstance(provider, GeminiProvider)
+
+    def test_gemini_raises_without_api_key(self):
+        """get_ai_provider raises ValueError when gemini_api_key is missing."""
+        with patch(
+            "app.utils.ai_provider.settings",
+            self._mock_settings(ai_provider="gemini", gemini_api_key=None),
+        ):
+            with pytest.raises(ValueError, match="GEMINI_API_KEY"):
+                get_ai_provider()
+
+    @patch("openai.OpenAI")
+    def test_returns_ollama_provider(self, mock_openai_cls):
+        """get_ai_provider returns OllamaProvider for ai_provider='ollama'."""
+        with patch(
+            "app.utils.ai_provider.settings",
+            self._mock_settings(ai_provider="ollama"),
+        ):
+            provider = get_ai_provider()
+        assert isinstance(provider, OllamaProvider)
+
+    @patch("openai.OpenAI")
+    def test_returns_openrouter_provider(self, mock_openai_cls):
+        """get_ai_provider returns OpenRouterProvider for ai_provider='openrouter'."""
+        with patch(
+            "app.utils.ai_provider.settings",
+            self._mock_settings(ai_provider="openrouter", openrouter_api_key="or-key"),
+        ):
+            provider = get_ai_provider()
+        assert isinstance(provider, OpenRouterProvider)
+
+    def test_openrouter_raises_without_api_key(self):
+        """get_ai_provider raises ValueError when openrouter_api_key is missing."""
+        with patch(
+            "app.utils.ai_provider.settings",
+            self._mock_settings(ai_provider="openrouter", openrouter_api_key=None),
+        ):
+            with pytest.raises(ValueError, match="OPENROUTER_API_KEY"):
+                get_ai_provider()
+
+    def test_returns_litellm_provider(self):
+        """get_ai_provider returns LiteLLMProvider for ai_provider='litellm'."""
+        with patch(
+            "app.utils.ai_provider.settings",
+            self._mock_settings(ai_provider="litellm"),
+        ):
+            provider = get_ai_provider()
+        assert isinstance(provider, LiteLLMProvider)
+
+    def test_raises_for_unknown_provider(self):
+        """get_ai_provider raises ValueError for an unrecognised provider name."""
+        with patch(
+            "app.utils.ai_provider.settings",
+            self._mock_settings(ai_provider="unsupported_provider"),
+        ):
+            with pytest.raises(ValueError, match="Unknown AI provider"):
+                get_ai_provider()
+
+    def test_provider_name_is_case_insensitive(self):
+        """get_ai_provider normalises provider names to lowercase."""
+        with patch(
+            "app.utils.ai_provider.settings",
+            self._mock_settings(ai_provider="anthropic", anthropic_api_key="ant-key"),
+        ):
+            # "Anthropic" and "ANTHROPIC" should work just like "anthropic"
+            provider = get_ai_provider()
+        assert isinstance(provider, AnthropicProvider)
+
+    def test_litellm_passes_non_default_base_url(self):
+        """LiteLLM provider receives api_base when openai_base_url differs from default."""
+        with patch(
+            "app.utils.ai_provider.settings",
+            self._mock_settings(
+                ai_provider="litellm",
+                openai_base_url="http://my-proxy/v1",
+            ),
+        ):
+            provider = get_ai_provider()
+        assert isinstance(provider, LiteLLMProvider)
+        assert provider._api_base == "http://my-proxy/v1"
+
+    def test_litellm_omits_base_url_when_default(self):
+        """LiteLLM provider has api_base=None when openai_base_url is the default."""
+        with patch(
+            "app.utils.ai_provider.settings",
+            self._mock_settings(
+                ai_provider="litellm",
+                openai_base_url="https://api.openai.com/v1",
+            ),
+        ):
+            provider = get_ai_provider()
+        assert isinstance(provider, LiteLLMProvider)
+        assert provider._api_base is None
diff --git a/tests/test_extract_metadata_gpt.py b/tests/test_extract_metadata_gpt.py
index a39325e0..ff96bad8 100644
--- a/tests/test_extract_metadata_gpt.py
+++ b/tests/test_extract_metadata_gpt.py
@@ -67,12 +67,11 @@ class TestExtractMetadataWithGpt:
 
     @patch("app.tasks.extract_metadata_with_gpt.embed_metadata_into_pdf")
     @patch("app.tasks.extract_metadata_with_gpt.log_task_progress")
-    @patch("app.tasks.extract_metadata_with_gpt.client")
-    def test_successful_metadata_extraction(self, mock_client, mock_log_progress, mock_embed_task):
-        """Test successful metadata extraction with valid GPT response."""
-        # Mock the OpenAI client response
-        mock_completion = MagicMock()
-        mock_completion.choices[0].message.content = json.dumps(
+    @patch("app.utils.ai_provider.get_ai_provider")
+    def test_successful_metadata_extraction(self, mock_get_provider, mock_log_progress, mock_embed_task):
+        """Test successful metadata extraction with valid AI provider response."""
+        mock_provider = MagicMock()
+        mock_provider.chat_completion.return_value = json.dumps(
             {
                 "filename": "2024-01-15_Invoice_Amazon",
                 "empfaenger": "John Doe",
@@ -89,7 +88,7 @@ class TestExtractMetadataWithGpt:
                 "monetary_amounts": ["99.99 EUR"],
             }
         )
-        mock_client.chat.completions.create.return_value = mock_completion
+        mock_get_provider.return_value = mock_provider
 
         # Set task request context directly on the Celery task
         extract_metadata_with_gpt.request.id = "test-task-id"
@@ -97,11 +96,11 @@ class TestExtractMetadataWithGpt:
         # Call the underlying function directly (not through Celery)
         result = extract_metadata_with_gpt.__wrapped__("test_invoice.pdf", "Invoice from Amazon for 99.99 EUR", 123)
 
-        # Verify OpenAI was called
-        mock_client.chat.completions.create.assert_called_once()
-        call_args = mock_client.chat.completions.create.call_args
-        assert call_args[1]["temperature"] == 0
-        assert len(call_args[1]["messages"]) == 2
+        # Verify AI provider was called
+        mock_provider.chat_completion.assert_called_once()
+        call_kwargs = mock_provider.chat_completion.call_args[1]
+        assert call_kwargs["temperature"] == 0
+        assert len(call_kwargs["messages"]) == 2
 
         # Verify metadata was extracted correctly
         assert result["s3_file"] == "test_invoice.pdf"
@@ -117,14 +116,14 @@ class TestExtractMetadataWithGpt:
 
     @patch("app.tasks.extract_metadata_with_gpt.embed_metadata_into_pdf")
     @patch("app.tasks.extract_metadata_with_gpt.log_task_progress")
-    @patch("app.tasks.extract_metadata_with_gpt.client")
-    def test_handles_json_in_backticks(self, mock_client, mock_log_progress, mock_embed_task):
+    @patch("app.utils.ai_provider.get_ai_provider")
+    def test_handles_json_in_backticks(self, mock_get_provider, mock_log_progress, mock_embed_task):
         """Test extraction handles JSON wrapped in markdown code blocks."""
-        mock_completion = MagicMock()
-        mock_completion.choices[
-            0
-        ].message.content = '```json\n{"filename": "test.pdf", "document_type": "Unknown"}\n```'
-        mock_client.chat.completions.create.return_value = mock_completion
+        mock_provider = MagicMock()
+        mock_provider.chat_completion.return_value = (
+            '```json\n{"filename": "test.pdf", "document_type": "Unknown"}\n```'
+        )
+        mock_get_provider.return_value = mock_provider
 
         extract_metadata_with_gpt.request.id = "test-task-id"
 
@@ -136,12 +135,12 @@ class TestExtractMetadataWithGpt:
 
     @patch("app.tasks.extract_metadata_with_gpt.embed_metadata_into_pdf")
     @patch("app.tasks.extract_metadata_with_gpt.log_task_progress")
-    @patch("app.tasks.extract_metadata_with_gpt.client")
-    def test_handles_invalid_json_response(self, mock_client, mock_log_progress, mock_embed_task):
-        """Test handling of invalid JSON in GPT response."""
-        mock_completion = MagicMock()
-        mock_completion.choices[0].message.content = "This is not valid JSON at all"
-        mock_client.chat.completions.create.return_value = mock_completion
+    @patch("app.utils.ai_provider.get_ai_provider")
+    def test_handles_invalid_json_response(self, mock_get_provider, mock_log_progress, mock_embed_task):
+        """Test handling of invalid JSON in AI provider response."""
+        mock_provider = MagicMock()
+        mock_provider.chat_completion.return_value = "This is not valid JSON at all"
+        mock_get_provider.return_value = mock_provider
 
         extract_metadata_with_gpt.request.id = "test-task-id"
 
@@ -155,10 +154,12 @@ class TestExtractMetadataWithGpt:
 
     @patch("app.tasks.extract_metadata_with_gpt.embed_metadata_into_pdf")
     @patch("app.tasks.extract_metadata_with_gpt.log_task_progress")
-    @patch("app.tasks.extract_metadata_with_gpt.client")
-    def test_handles_openai_api_exception(self, mock_client, mock_log_progress, mock_embed_task):
-        """Test handling of OpenAI API exceptions."""
-        mock_client.chat.completions.create.side_effect = Exception("API Error: Rate limit exceeded")
+    @patch("app.utils.ai_provider.get_ai_provider")
+    def test_handles_api_exception(self, mock_get_provider, mock_log_progress, mock_embed_task):
+        """Test handling of AI provider API exceptions."""
+        mock_provider = MagicMock()
+        mock_provider.chat_completion.side_effect = Exception("API Error: Rate limit exceeded")
+        mock_get_provider.return_value = mock_provider
 
         extract_metadata_with_gpt.request.id = "test-task-id"
 
@@ -172,15 +173,15 @@ class TestExtractMetadataWithGpt:
 
     @patch("app.tasks.extract_metadata_with_gpt.embed_metadata_into_pdf")
     @patch("app.tasks.extract_metadata_with_gpt.log_task_progress")
-    @patch("app.tasks.extract_metadata_with_gpt.client")
+    @patch("app.utils.ai_provider.get_ai_provider")
     @patch("app.tasks.extract_metadata_with_gpt.SessionLocal")
     def test_retrieves_file_id_from_database_when_not_provided(
-        self, mock_session_local, mock_client, mock_log_progress, mock_embed_task
+        self, mock_session_local, mock_get_provider, mock_log_progress, mock_embed_task
     ):
         """Test file_id retrieval from database when not provided."""
-        mock_completion = MagicMock()
-        mock_completion.choices[0].message.content = '{"filename": "test.pdf", "document_type": "Unknown"}'
-        mock_client.chat.completions.create.return_value = mock_completion
+        mock_provider = MagicMock()
+        mock_provider.chat_completion.return_value = '{"filename": "test.pdf", "document_type": "Unknown"}'
+        mock_get_provider.return_value = mock_provider
 
         # Mock database session
         mock_db = MagicMock()
@@ -206,15 +207,15 @@ class TestExtractMetadataWithGpt:
 
     @patch("app.tasks.extract_metadata_with_gpt.embed_metadata_into_pdf")
     @patch("app.tasks.extract_metadata_with_gpt.log_task_progress")
-    @patch("app.tasks.extract_metadata_with_gpt.client")
-    def test_validates_filename_security(self, mock_client, mock_log_progress, mock_embed_task):
+    @patch("app.utils.ai_provider.get_ai_provider")
+    def test_validates_filename_security(self, mock_get_provider, mock_log_progress, mock_embed_task):
         """Test filename validation to prevent path traversal."""
-        mock_completion = MagicMock()
+        mock_provider = MagicMock()
         # Try to inject a malicious filename
-        mock_completion.choices[0].message.content = json.dumps(
+        mock_provider.chat_completion.return_value = json.dumps(
             {"filename": "../../../etc/passwd", "document_type": "Invoice"}
         )
-        mock_client.chat.completions.create.return_value = mock_completion
+        mock_get_provider.return_value = mock_provider
 
         extract_metadata_with_gpt.request.id = "test-task-id"
 
@@ -226,14 +227,14 @@ class TestExtractMetadataWithGpt:
 
     @patch("app.tasks.extract_metadata_with_gpt.embed_metadata_into_pdf")
     @patch("app.tasks.extract_metadata_with_gpt.log_task_progress")
-    @patch("app.tasks.extract_metadata_with_gpt.client")
-    def test_validates_filename_with_dots(self, mock_client, mock_log_progress, mock_embed_task):
+    @patch("app.utils.ai_provider.get_ai_provider")
+    def test_validates_filename_with_dots(self, mock_get_provider, mock_log_progress, mock_embed_task):
         """Test filename validation rejects '..' in filenames."""
-        mock_completion = MagicMock()
-        mock_completion.choices[0].message.content = json.dumps(
+        mock_provider = MagicMock()
+        mock_provider.chat_completion.return_value = json.dumps(
             {"filename": "test..invoice.pdf", "document_type": "Invoice"}
         )
-        mock_client.chat.completions.create.return_value = mock_completion
+        mock_get_provider.return_value = mock_provider
 
         extract_metadata_with_gpt.request.id = "test-task-id"
 
@@ -244,14 +245,14 @@ class TestExtractMetadataWithGpt:
 
     @patch("app.tasks.extract_metadata_with_gpt.embed_metadata_into_pdf")
     @patch("app.tasks.extract_metadata_with_gpt.log_task_progress")
-    @patch("app.tasks.extract_metadata_with_gpt.client")
-    def test_accepts_valid_filename(self, mock_client, mock_log_progress, mock_embed_task):
+    @patch("app.utils.ai_provider.get_ai_provider")
+    def test_accepts_valid_filename(self, mock_get_provider, mock_log_progress, mock_embed_task):
         """Test that valid filenames are accepted."""
-        mock_completion = MagicMock()
-        mock_completion.choices[0].message.content = json.dumps(
+        mock_provider = MagicMock()
+        mock_provider.chat_completion.return_value = json.dumps(
             {"filename": "2024-01-15_Invoice_Amazon.pdf", "document_type": "Invoice"}
         )
-        mock_client.chat.completions.create.return_value = mock_completion
+        mock_get_provider.return_value = mock_provider
 
         extract_metadata_with_gpt.request.id = "test-task-id"
 
@@ -262,12 +263,12 @@ class TestExtractMetadataWithGpt:
 
     @patch("app.tasks.extract_metadata_with_gpt.embed_metadata_into_pdf")
     @patch("app.tasks.extract_metadata_with_gpt.log_task_progress")
-    @patch("app.tasks.extract_metadata_with_gpt.client")
-    def test_handles_malformed_json_with_valid_structure(self, mock_client, mock_log_progress, mock_embed_task):
+    @patch("app.utils.ai_provider.get_ai_provider")
+    def test_handles_malformed_json_with_valid_structure(self, mock_get_provider, mock_log_progress, mock_embed_task):
         """Test handling of JSON that's parseable but missing expected fields."""
-        mock_completion = MagicMock()
-        mock_completion.choices[0].message.content = '{"unexpected_field": "value"}'
-        mock_client.chat.completions.create.return_value = mock_completion
+        mock_provider = MagicMock()
+        mock_provider.chat_completion.return_value = '{"unexpected_field": "value"}'
+        mock_get_provider.return_value = mock_provider
 
         extract_metadata_with_gpt.request.id = "test-task-id"
 
@@ -279,12 +280,12 @@ class TestExtractMetadataWithGpt:
 
     @patch("app.tasks.extract_metadata_with_gpt.embed_metadata_into_pdf")
     @patch("app.tasks.extract_metadata_with_gpt.log_task_progress")
-    @patch("app.tasks.extract_metadata_with_gpt.client")
-    def test_handles_absolute_path_filename(self, mock_client, mock_log_progress, mock_embed_task):
-        """Test handling when filename is provided as an absolute path (line 73)."""
-        mock_completion = MagicMock()
-        mock_completion.choices[0].message.content = '{"filename": "test.pdf", "document_type": "Unknown"}'
-        mock_client.chat.completions.create.return_value = mock_completion
+    @patch("app.utils.ai_provider.get_ai_provider")
+    def test_handles_absolute_path_filename(self, mock_get_provider, mock_log_progress, mock_embed_task):
+        """Test handling when filename is provided as an absolute path."""
+        mock_provider = MagicMock()
+        mock_provider.chat_completion.return_value = '{"filename": "test.pdf", "document_type": "Unknown"}'
+        mock_get_provider.return_value = mock_provider
 
         extract_metadata_with_gpt.request.id = "test-task-id"
 
@@ -298,15 +299,15 @@ class TestExtractMetadataWithGpt:
 
     @patch("app.tasks.extract_metadata_with_gpt.embed_metadata_into_pdf")
     @patch("app.tasks.extract_metadata_with_gpt.log_task_progress")
-    @patch("app.tasks.extract_metadata_with_gpt.client")
+    @patch("app.utils.ai_provider.get_ai_provider")
     @patch("app.tasks.extract_metadata_with_gpt.SessionLocal")
     def test_database_lookup_with_existing_file(
-        self, mock_session_local, mock_client, mock_log_progress, mock_embed_task
+        self, mock_session_local, mock_get_provider, mock_log_progress, mock_embed_task
     ):
-        """Test file_id retrieval when file exists on disk and in database (branches 76->82, 79->82)."""
-        mock_completion = MagicMock()
-        mock_completion.choices[0].message.content = '{"filename": "test.pdf", "document_type": "Unknown"}'
-        mock_client.chat.completions.create.return_value = mock_completion
+        """Test file_id retrieval when file exists on disk and in database."""
+        mock_provider = MagicMock()
+        mock_provider.chat_completion.return_value = '{"filename": "test.pdf", "document_type": "Unknown"}'
+        mock_get_provider.return_value = mock_provider
 
         # Mock database session
         mock_db = MagicMock()
@@ -333,20 +334,14 @@ class TestExtractMetadataWithGpt:
 
 
 @pytest.mark.unit
-class TestClientInitialization:
-    """Tests for OpenAI client initialization error handling."""
+class TestModuleImports:
+    """Tests for module import behaviour."""
 
-    def test_client_initialization_imports_successfully(self):
-        """Test that module imports successfully even if client initialization fails (lines 25-27).
+    def test_module_imports_successfully(self):
+        """Test that the module imports successfully."""
+        import app.tasks.extract_metadata_with_gpt as mod
 
-        The module has a try/except block for client initialization that sets client to None
-        on failure. This test verifies the module can be imported without crashing,
-        regardless of whether the client initializes successfully or not.
-        """
-        # Import should succeed regardless of client initialization success
-        from app.tasks.extract_metadata_with_gpt import client
-
-        # Client will be either an OpenAI client instance or None
-        # Both are valid states - the important thing is the import doesn't crash
-        # We verify the client variable exists and has a defined type
-        assert hasattr(client, "__class__") or client is None
+        assert mod is not None
+        assert hasattr(mod, "extract_metadata_with_gpt")
+        assert hasattr(mod, "extract_json_from_text")
+        assert hasattr(mod, "get_ai_provider")
diff --git a/tests/test_ocr_processing.py b/tests/test_ocr_processing.py
index 29d3af20..c2c9159e 100644
--- a/tests/test_ocr_processing.py
+++ b/tests/test_ocr_processing.py
@@ -407,65 +407,60 @@ startxref
 
 @pytest.mark.unit
 class TestRefineTextWithGPT:
-    """Tests for OpenAI text refinement task."""
+    """Tests for AI provider text refinement task."""
 
     @patch("app.tasks.refine_text_with_gpt.log_task_progress")
     def test_successful_text_refinement(self, mock_log):
-        """Test successful text refinement with OpenAI."""
+        """Test successful text refinement with AI provider."""
         raw_text = "This is s0me text with OCR err0rs"
         filename = "test.pdf"
-
-        # Mock OpenAI response
-        mock_choice = Mock()
-        mock_choice.message.content = "This is some text with OCR errors"
-
-        mock_response = Mock()
-        mock_response.choices = [mock_choice]
-
-        mock_client = Mock()
-        mock_client.chat.completions.create.return_value = mock_response
+        cleaned = "This is some text with OCR errors"
 
         # Import the module to patch the correct function
         from app.tasks import extract_metadata_with_gpt as metadata_module
 
+        mock_provider = MagicMock()
+        mock_provider.chat_completion.return_value = cleaned
+
         with (
-            patch("app.tasks.refine_text_with_gpt.client", mock_client),
+            patch("app.utils.ai_provider.get_ai_provider", return_value=mock_provider),
             patch.object(metadata_module, "extract_metadata_with_gpt") as mock_extract,
             patch("app.tasks.refine_text_with_gpt.settings") as mock_settings,
         ):
             mock_settings.openai_model = "gpt-4"
+            mock_settings.ai_model = None
             mock_extract.delay = MagicMock()
 
             result = refine_text_with_gpt.run(filename, raw_text)
 
             # Verify results
             assert result["filename"] == filename
-            assert result["cleaned_text"] == "This is some text with OCR errors"
+            assert result["cleaned_text"] == cleaned
 
-            # Verify OpenAI was called correctly
-            mock_client.chat.completions.create.assert_called_once()
-            call_kwargs = mock_client.chat.completions.create.call_args[1]
-            assert call_kwargs["model"] == "gpt-4"
+            # Verify AI provider was called correctly
+            mock_provider.chat_completion.assert_called_once()
+            call_kwargs = mock_provider.chat_completion.call_args[1]
             assert len(call_kwargs["messages"]) == 2
             assert call_kwargs["messages"][1]["content"] == raw_text
 
             # Verify metadata extraction was queued
-            mock_extract.delay.assert_called_once_with(filename, "This is some text with OCR errors")
+            mock_extract.delay.assert_called_once_with(filename, cleaned)
 
     @patch("app.tasks.refine_text_with_gpt.log_task_progress")
     def test_openai_api_error(self, mock_log):
-        """Test error handling when OpenAI API fails."""
+        """Test error handling when AI provider call fails."""
         raw_text = "Test text"
         filename = "test.pdf"
 
-        mock_client = Mock()
-        mock_client.chat.completions.create.side_effect = Exception("OpenAI API error")
+        mock_provider = MagicMock()
+        mock_provider.chat_completion.side_effect = Exception("OpenAI API error")
 
         with (
-            patch("app.tasks.refine_text_with_gpt.client", mock_client),
+            patch("app.utils.ai_provider.get_ai_provider", return_value=mock_provider),
             patch("app.tasks.refine_text_with_gpt.settings") as mock_settings,
         ):
             mock_settings.openai_model = "gpt-4"
+            mock_settings.ai_model = None
 
             # Should raise the exception
             with pytest.raises(Exception) as exc_info: