Build a Databricks App OAuth config from an agent's ``litellm_params``. Returns ``None`` when the agent has no ``databricks_oauth`` block. Raises

Perform health check with shared state coordination. This method: 1. First checks if there are recent cached results

Return if request can be processed. Returns: - True: * If healthy deployments are available * OR If

Process input prompt by applying guardrails to text content. Args: data: Request data dictionary containing 'prompt' par

Process the input for multimodal embedding requests. Args: _input (Union[list, str]): The input data to process.

Process output transcription by applying guardrails to transcribed text. Args: response: Transcription response object c

Create a copy of headers with sensitive values redacted for safe logging. Shows first 8 characters of sensitive values for debugging

Register a guardrail for onboarding (team submission). Accepts a guardrail config in the [Generic Guardrail API](https://docs.litellm.ai

Resets the budget for all the litellm keys Catches Exceptions and logs them

For keys and teams with budget_limits, reset any individual windows where reset_at <= now. Only the expired windows are reset; other

Resolve model_name from model_id. This method attempts to find the correct model_name to use with the router so that litellm

Return a copy of *operation* with fully-resolved, merged parameters. Handles two common patterns in real-world OpenAPI specs: 1. **$ref para

Map a pass-through ``custom_llm_provider`` to the provider scope that namespaces passthrough managed object IDs, or ``None`` when the route is not

Return an OAuth2 credential whose access_token is good for the next request. Returns the credential unchanged while its token is valid for at lea

- assert if response headers < 4kb (nginx limit).

Map an OpenAPI operationId / fallback to a provider-safe tool name. Replaces any character outside ``[a-zA-Z0-9_-]`` with ``_`` and caps the

Anchor the request's root (server) span for explicit child parenting. No-ops for a non-recordable span so a bad capture can never replace a good

Add the most accurate rate limit headers for a given model response. ## TODO: add model group rate limit headers # - if heal

Split text into chunks at word boundaries without breaking words. Always returns at least one chunk. Short text (≤ max_length) is

Mark the worker as draining. Idempotent — repeated calls (e.g. SIGTERM followed by a preStop hit on /health/drain) do not reset the c

Return a new message list with empty or whitespace-only ``{"type": "text"}`` content blocks removed. Anthropic's API rejects requests co

Transform file content response.

Transform the overall request to be sent to the API. Returns: dict: The transformed request. Sent as the body of the API

Transform provider-specific batch retrieval response to LiteLLM format. Args: model: Model name raw_response

Translate Anthropic's output_format to OpenAI's response_format. Anthropic output_format: {"type": "json_schema", "schema": {...}}

Transform streaming litellm completion chunk to Google GenAI generate_content format Args: response: Streaming ModelResp

Convert Anthropic tool definitions to Responses API function tools.

Return a copy of *messages* with long base64 data-URI payloads replaced by human-readable size placeholders.

Confirm the tool choice is passed in the OpenAI format. Prevents user errors like: https://github.com/BerriAI/litellm/issues/7483

Checks if the environment variables are valid for the given model. Args: model (Optional[str]): The name of the model. Defaults to N

Validate that a response from litellm.responses() or litellm.aresponses() conforms to the expected ResponsesAPIResponse structure. Args:

Watsonx pass-through endpoint. Allows using Watsonx APIs with automatic IAM token management and version parameter injection. Example:

Extract cache_control from source and add to target if it should be preserved. This method accepts Any type to support both regular

Adds router settings from DB config to litellm proxy 1. Get router settings from DB 2. Get router settings from config

Helper function to make a generic LLM API call through the router, this allows you to use retries/fallbacks with litellm router

Build a single PartType for an input element, handling text, data URIs, file references, and GCS URLs.