Returns -> Deployment or None Raise Exception -> if model found in invalid format Optimized with O(1) index lookup instead

Get the global PolicyRegistry singleton. Returns: The global PolicyRegistry instance

Helper to checks if provided route is an OpenAI route Returns: - True: if route is an OpenAI route - False:

Build a minimal kwargs dict representative of a litellm callback invocation.

Common helper to route the request

The same as `litellm.litellm_core_utils.token_counter`. Kept for backwards compatibility.

Accept ``redirect_uri`` when it is (a) same-origin with the proxy's own request origin, (b) loopback, (c) listed in the ``MCP_TRUSTED_REDIRECT

Create a new credential. Args: credential_name (str): Name of the credential credential_info (Dict[str, Any]

Pseudo-destructor to be invoked to clean up global data structures when router is no longer used. For now, unhook router's callbacks

Return a urlsafe-base64 managed ID string (trailing ``=`` stripped).

Transform the image generation request to Gemini format For Gemini 2.5 Flash Image Preview, use the standard Gemini format with resp

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

Get list of allowed tool names for a specific server based on key/team permissions. Follows same inheritance logic as get_allowed_mcp

Resolve the model(s) a request targets, for model-access and budget checks. Returns ``None`` when the request was dispatched to a user-defined pa

Converts a string to a boolean if it's a recognized boolean string. Returns None if the string is not a recognized boolean value. :param

Use `/team/member_add` AND `/team/member/delete` to add/remove new team members You can now update team budget / rate limits via /team/updat

Whether ``model`` uses adaptive thinking (``output_config.effort``). The model cost map is authoritative: an explicit ``supports_adaptive_thi

Returns -> Deployment or None Raise Exception -> if model found in invalid format

Returns the provider config for a given provider. Uses O(1) dictionary lookup for fast provider resolution. Python classes t

Get a proxy hook from the proxy_hook_mapping

Get the registered MCP Servers from the registry and union with the config MCP Servers

Handle MCP requests through StreamableHTTP.

Mint managed IDs for raw provider values listed in ``BUILTIN_OUTPUT_ID_FIELD_MAP`` and swap them into *body*. Returns the same *body* ob

Validate that MCP servers requested on a key are within the allowed scope. Rules: - If key is in a team: key's mcp_servers must be a sub

for cloudflare ai gateway - https://github.com/BerriAI/litellm/issues/4317 Handles custom api_base for: 1. Gemini (Google AI

Make sure each key starts with the given namespace

Filter out unsupported fields from JSON schema for Anthropic's output_format API. Anthropic's output_format doesn't support certain

Get the next standardized reset time based on the duration. All durations will reset at predictable intervals, aligned from the current time

Invoke an agent using the A2A protocol (JSON-RPC 2.0). Supported methods: - message/send: Send a message and get a response - messag

Transform the image generation response to the litellm image response

Update a new mcp server record in the db

Moderate prompts (request) and responses (response); fail-open. - ``request``: evaluate the prompt via the before_prompt webhook and