Async implementation of '_get_cooldown_deployments'

Match production entrypoint: ``_run_deferred_stream_guardrails`` uses dispatch.

Classify a Cursor API request into a readable operation name.

Convert consecutive user messages at the end to guarded_text type if guardrailConfig is present and no guarded_text is already presen

Convert MCP SamplingMessage list to OpenAI messages format. MCP messages use: - role: "user" | "assistant" - content: TextContent | I

Create appropriate Bedrock content block based on mime type.

Decrypts a dictionary of variables and returns them.

Enforce upperbound limits on key parameters. For key generation (fill_defaults=True): fills None values with upperbound defaults. For ke

Extract all messages (request + response) with sequence numbers and timestamps. Processes request messages from StandardLoggingPaylo

Extract text content from the last user message for routing. Handles tool-call conversations (where the last message may be an

Fire the deferred streaming logging callback after the full streaming pipeline (including guardrail end-of-stream blocks) has complet

Generic function that handles lazy importing for most attributes. This is the workhorse function - it does the actual importing and caching.

Return all deployments of a model name Used for accurate 'get_model_list'. if team_id specified, only return team-specific

Get allowed MCP servers for a team. Unions two sources: - Legacy team.object_permission (mcp_servers, mcp_access_groups,

Get common email parameters used across different email sending methods Args: email_event: Type of email event

Create an OpenTelemetry Resource using config-driven defaults.

Handle logging for proxy only errors by calling `litellm_logging_obj.async_failure_handler` Is triggered when self._is_proxy_only_er

Handle ``GetModelInvocationJob`` status check for AWS Bedrock bulk batch inference jobs (the ARN type returned by ``CreateModelInvoca

Initialize a custom logger compatible class

Return a mock prisma_client with litellm_tooltable.upsert, find_many, find_unique.

Minimal fake Starlette Request.

For GET/DELETE endpoints that cannot carry a JSON body, read a JSON-encoded ``litellm_params_template`` query parameter and merge its con

Whether a route prefix (e.g. ``mantle/``) appears as a leading path segment. A route token is only valid at the start of the model id or imme

Distinct body.model values, mirroring how the rate limiter collects the models from a streamed batch file before the access check.

Parse S3 URI to extract bucket name and object key. Args: s3_uri: S3 URI (e.g., "s3://bucket-name/path/to/file")

Filter out model in model group, if: - model context window < message length. For azure openai models, requires 'base_model' is set.

Raises an HTTPException if the user is not a premium user

Helper function to prepare MCP server data for database operations. Handles JSON field serialization for mcp_info and env fields. Args:

Replace credential-bearing values in ``litellm_params`` with ``REDACTED_BY_LITELM`` while preserving non-secret keys (``api_base``, ``reg

Resolve an LLM model name from MCP ModelPreferences. Strategy: 1. Check hints for substring matches against known model names. 2. Fal

Compare request bodies as bytes; never invokes ``json.loads``. vcrpy's stock ``body`` matcher unconditionally json-decodes ``application/json

Skip downloading batch input files when the operator disabled batch input-file rate limiting, when the batch runs entirely on a skip-

Strip markdown code block wrapper from JSON content if present. SAP GenAI Hub with Anthropic models sometimes returns JSON wrapped in

Strip fields not permitted by FireworksAI from messages.

Transforms the response from the Bedrock embedding provider to the OpenAI format.

Validate if user has permission to add members to the team. Standard users can self-join an *available team*, but the bypass must not be allo

Raise HTTPException 403 if any validation rule doesn't match the token response. Supports dot-notation for nested fields (e.g. ``"team.enterprise

Send text to the Presidio analyzer endpoint and get analysis results

Merge ``x-litellm-tags`` header tags into ``request_data`` BEFORE auth budget gates run, so ``_tag_max_budget_check`` (which only

Iterate through all models in a model group to check for batch Future Improvement - cache the result.

Async: Send a message to an A2A agent. Uses the @client decorator for LiteLLM logging and tracking. If litellm_params contains custom_ll

Key - the key in cache Value - float - the value you want to increment by Refresh_ttl - bool - if True, resets the Redis TT

Async function to write a secret to AWS Secrets Manager Args: secret_name: Name of the secret secret_value:

Exchange an authorization code for a short-lived BYOK session JWT. 1. Validates the authorization code and PKCE challenge. 2. Stores the

Call an MCP Tool.

Call a tool with the given name and arguments Args: server_name: Server name name: Tool name arg

Helper function to create a mock router with fallback configurations.

Decrypt ``scope="global"`` env var values in place after reading the DB. Accepts ``MCPEnvVar`` models (``LiteLLM_MCPServerTable``) or plain dicts

Exchange *subject_token* for a scoped access token. Returns the exchanged ``access_token`` string (suitable for a ``Bearer`` header).

Drop detections that fall below configured per-entity score thresholds or match an entity type in the deny list.

Find and validate specific team ID from team_id_jwt_field or team_alias_jwt_field

Get an agent by its ID from the database

Get the AWS region name for non-llm api calls. LLM API calls check the model arn and end up using that as the region name.

Get Azure AD token provider based on Service Principal with Secret workflow. Based on: https://github.com/openai/openai-python/blob/main/exa

Extract the base prompt ID by stripping the version suffix if present. Args: prompt_id: Prompt ID that may include version suffix (e