Encode provider and model_id into video_id using base64.

Get the `custom_llm_provider` from the request header `custom-llm-provider`

Retrieves a value from a nested dictionary using dot notation. Args: data: The dictionary to search in key_path: The path to

Sets a span attribute safely with OTEL-compliant primitive typing for Arize/Phoenix.

Encode a value for DualCache / Redis (``json.dumps``-safe). If ``model_type`` is set, the payload is validated with that model, then

Build filter conditions for key listing. Visibility rules: - Users always see their own keys (user_id match) - Team admins see ALL keys f

Get the AWS region name from the environment variables. Parameters: optional_params (dict): Optional parameters for the

Build a Prisma-like row object.

Check if the given model supports function calling and return a boolean value. Parameters: model (str): The model name to be checked.

Transforms OpenAI JSONL content to Bedrock batch format Bedrock batch format: { "recordId": "alphanumeric string", "modelInput": {JS

Add a success callback to `litellm.success_callback`. Auto-routes async callbacks to litellm._async_success_callback. Special

Count records matching the criteria.

Get the `custom_llm_provider` from the request query parameters Safely reads the request query parameters

Get API credentials and provider info from a model name in model_list. Useful for passthrough endpoints (files, batches, etc.) that n

Get the labels for a metric, filtered if configured. The result is cached on the instance so the label set used to construct

Get the MCP Server from the server id

True if the call was rejected for being over budget (vs a provider error).

Load attachments from a configuration list. Args: attachments_config: List of attachment dictionaries from YAML.

Performs the actual redaction on the logging object and result. redact_streaming_responses=False skips the in-place redaction of the shared

Allow user to modify outgoing data Covers: 1. /chat/completions 2. /embeddings 3. /image/generation

Process input messages by applying guardrails to text content.

Safely get Vertex AI project without mutating the litellm_params dict. Unlike get_vertex_ai_project(), this does NOT pop values from

Construct the handler with auth + proxy seams patched at the instance level. ``_ensure_access_token`` and ``_check_custom_proxy`` are inherited f

Map OpenAI-style tools/functions to Vertex AI format. Args: value: List of tool definitions optional_params:

Build the minimal proxy_server module attributes that _run_centralized_common_checks reads. ``master_key`` defaults to a non-None value becau

Run ``common_checks`` once at the ``user_api_key_auth`` wrapper boundary, regardless of which ``_user_api_key_auth_builder`` path returned. Th

Accept the WebSocket connection

Add server name prefix to any MCP resource name.

Create an audit log for an object.

Get the MCP Server from the server name. Uses priority-based matching to avoid collisions: 1. First pass: exact alias match

if 'model_name' is none, returns all. Returns list of model id's. Optimized with O(1) or O(k) index lookup when model_name

Guarantees response from 'get_secret' is either boolean or none. Used for fixing linting errors. Args: secret_name: The name of the

No global setup needed today; present for lifecycle symmetry.

Setup guardrail with all Canadian PII patterns (mimics the policy template)

Check if the given model supports function calling and return a boolean value. Parameters: model (str): The model name to be checked.

Keep this state less - leave the state management (e.g. tracking current_output_item_id, current_response_id, current_conversation_id, curren

OpenAI tools looks like: tools = [ { "type": "function", "function": { "name": "get_current_w

Inspect ``litellm.callbacks`` once and answer the per-hook capability questions used to short-circuit no-op work on the chat-completi

Build a mock httpx response with the given JSON body.

Prisma mock where only the given ids resolve to a user row / team row. ``teams`` are teams a SCIM group write created, so they carry provenance;

Process patch operations for a group and return update data, final members and, when the request contained a member ``replace`` op, the absolute t

Transforms ResponseAPIUsage or ImageUsage to a Usage object. Both have the same spec with input_tokens, output_tokens, and i