Adds/updates to kwargs: - num_retries - litellm_trace_id - metadata

Convert an OpenAI-format message array to OCI GENERIC format.

Implementing async callbacks, to handle asyncio event loop issues when custom integrations need to use async functions.

Raise HTTP 403 if the user's role is blocked from accessing the given feature by the UI settings stored in general_settings. Args:

Create an audit log for an object.

Create the rate limit keys for the given key and value.

Get the bedrock route for the given model.

Read current spend from the cross-pod spend counter. Reads Redis FIRST (authoritative cross-pod value), not DualCache's async_get_cache

Check if key valid

Resolve the request route from the ASGI scope, with ``root_path`` stripped. Prefer this over ``request.url.path`` for any auth, ACL, routing

Use this to create a new INTERNAL user with a budget. Internal Users can access LiteLLM Admin UI to make keys, request access to models.

Alerting based on thresholds: - https://github.com/BerriAI/litellm/issues/1298 - Responses taking too long - Requests are ha

Helper function to know if a provider implementation supports httpx timeout

Build JWT claims for the outbound MCP access token. Args: user_api_key_dict: LiteLLM auth context for the current reques

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

Call review_gate with safe defaults for the injectable hooks.

Returns the appropriate client based on the given deployment, kwargs, and client_type. Parameters: deployment (dict): Th

Lightweight server stand-in for the per-user env-var endpoints. The handlers only read ``server_id``/``server_name``/``alias``/``env_vars``/

Build a Prisma-like row object.

Common transformation logic across sync + async Gemini /generateContent calls.

Helper function to update all metadata fields (both premium and standard). Args: updated_kv: The key-value dict being used for the u

Arize is used mainly for LLM I/O tracing, sending router+caching metrics adds bloat to arize logs

Call from ``pytest_terminal_summary``. No-op on xdist workers.

Check if a record exists.

Handle MCP requests through StreamableHTTP.

Check if a callback is disabled via the x-litellm-disable-callbacks header or via `litellm_disabled_callbacks` in standard_callback_dynam

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

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

Returns True if the guardrail should be run on the event_type

No transformation is needed for Anthropic messages This takes in a request in the Anthropic /v1/messages API spec -> transforms it

Translate a full Anthropic /v1/messages request dict to litellm.responses() / litellm.aresponses() kwargs.

Check if the model has guardrails defined and merge them with existing guardrails in the request data. Args: data: The request data

Internal function. Returns the token and url for the call. Handles logic if it's google ai studio vs. vertex ai. Returns

Helper method to fetch tools from MCP servers based on server filtering criteria. Args: user_api_key_auth: User authenti

Whether ``model`` uses adaptive thinking (``output_config.effort``). Sourced solely from the model cost map's ``supports_adaptive_thinking``

Build a stub httpx-like response.

Strip echoed request input from provider error strings. Provider validation errors (e.g. OpenAI ``RateLimitError`` carrying 178 pydantic

Example Usage: ``` from litellm import Router client = Router(model_list = [ { "model_na

Async implementation of 'get_available_deployments'. Allows all cache calls to be made async => 10x perf impact (8rps -> 100 rps).

Check if user has permission to modify (delete/regenerate) a verification token. Rules: - Proxy admin can modify any key - Internal

Create a tool function for an OpenAPI operation. This function creates an async tool function that can be called with keyword arguments. Para

Get the base model from the given model name. Handle model names like: - "us.meta.llama3-2-11b-instruct-v1:0" -> "meta.llama3-2-11b-inst

Get all custom loggers that are instances of the given class type