MCPcopy Create free account

hub / github.com/BerriAI/litellm / functions

Functions202,479 in github.com/BerriAI/litellm

↓ 5 callersFunctionnative_count_request
(request: Mapping[str, object], model: str, policy: DropPolicy)
litellm/llms/openai/responses/count_tokens/transformation.py:406
↓ 5 callersFunctionnative_reply
(summary: str = "Project code MAPLE-47", signed: bool = True, truncated: bool = False)
tests/unit/router_strategy/complexity_router/test_context_compaction.py:116
↓ 5 callersMethodnative_result
(self, pending: NativePending[_T])
litellm/caching/redis_batch.py:597
↓ 5 callersFunctionnative_route
(*routes: Route)
tests/test_litellm_rust/support/routing.py:20
↓ 5 callersFunctionnd
(e,t)
litellm/proxy/_experimental/out/_next/static/chunks/14v3irtjifjt2.js:1
↓ 5 callersMethodneeds_iterator_wrap
Whether ``async_data_generator`` needs to wrap the upstream stream through ``async_post_call_streaming_iterator_hook``. Instance method
litellm/proxy/utils.py:3784
↓ 5 callersFunctionnewFallbackTestResourceData
(t *testing.T, model string, fallbackModels []interface{}, fallbackType string)
terraform/provider/litellm/resource_fallback_test.go:13
↓ 5 callersFunctionnew_project
Create a new project. Projects sit between teams and keys in the hierarchy. Only admins or team admins can create projects. # Parameter
enterprise/litellm_enterprise/proxy/management_endpoints/project_endpoints.py:462
↓ 5 callersFunctionnewest_marker
(text: str)
tests/integration/_support/responses_vendor.py:39
↓ 5 callersFunctionnewest_version_at_least
(packument: Packument, *, min_age: timedelta, as_of: datetime)
tests/e2e/cli_harness/npm_version_resolver.py:42
↓ 5 callersFunctionnn
(...e)
litellm/proxy/_experimental/out/_next/static/chunks/14v3irtjifjt2.js:1
↓ 5 callersMethodno_max_tokens_output_floor
Output-budget floor used when the request omits max_tokens. Capped at a fraction of the smallest configured TPM limit so a small per-
litellm/proxy/hooks/parallel_request_limiter_v3.py:788
↓ 5 callersMethodnormalize
(params: dict[str, object])
litellm/proxy/management_endpoints/cache_settings_endpoints.py:353
↓ 5 callersFunctionnormalize_agent_card
Convert an extended agent card to ``target``. When lowering to 0.3, ``additionalInterfaces`` is stripped so the conversion never re-exposes u
litellm/proxy/a2a/version_convert.py:118
↓ 5 callersFunctionnormalize_classifier_config_aliases
(config: Mapping[str, object])
litellm/router_strategy/complexity_router/config.py:694
↓ 5 callersFunctionnormalize_duration
(duration: str)
litellm/litellm_core_utils/duration_parser.py:25
↓ 5 callersFunctionnormalize_response
( model: &str, response: CohereResponse, )
litellm-rust/crates/llms/src/cohere/ocr/transformation.rs:238
↓ 5 callersFunctionnormalize_stream_event
Convert a single streamed JSON-RPC event's ``result`` to ``target``.
litellm/proxy/a2a/version_convert.py:74
↓ 5 callersFunctionnormalize_transcription_language_to_bcp47
OpenAI's transcription `language` param accepts bare ISO-639-1 codes like ``en``; speech APIs such as Google Speech-to-Text and NVIDIA Riva r
litellm/litellm_core_utils/audio_utils/utils.py:162
↓ 5 callersFunctionnormalized_cost
Map a raw $/1k-token cost into [0, 1] where 0 = most expensive, 1 = cheapest. Returns 0.5 when there's no spread.
litellm/router_strategy/adaptive_router/bandit.py:84
↓ 5 callersFunctionnormalized_metadata_slot
Return the request's metadata slot as a dict, normalising it in place first. Metadata can arrive as a JSON string (multipart/form-data, ``extra_b
litellm/proxy/litellm_pre_call_utils.py:650
↓ 5 callersFunctionnumeric_form_fields
The numeric fields of a request schema, mapped to the scalar to parse them as. Only a bare ``int``/``float`` or an optional one qualifies, s
litellm/proxy/common_utils/http_parsing_utils.py:149
↓ 5 callersFunctiono
()
litellm/proxy/_experimental/out/_next/static/chunks/1a2lzi3ursud4.js:1
↓ 5 callersFunctiono
(e,o,l)
litellm/proxy/_experimental/out/_next/static/chunks/2vx7-vp9_2q3a.js:1
↓ 5 callersFunctiono
(e,s,t,i)
litellm/proxy/_experimental/out/_next/static/chunks/2jxpy6eth1zvy.js:1
↓ 5 callersFunctionoD
(e,t)
litellm/proxy/_experimental/out/_next/static/chunks/0_6jqgnorqxl2.js:1
↓ 5 callersFunctionoauth_client_registration_matches
( registered_issuer: str | None, registered_url: str | None, current_issuer: str | None, curre
litellm/proxy/mcp_server/oauth_utils.py:747
↓ 5 callersFunctionoauth_config
(provider: Provider)
litellm/proxy/roi_calculator/oauth.py:75
↓ 5 callersMethodobserve_backend_event
(self, event: Mapping[str, object])
litellm/llms/xai/realtime/usage_meter.py:103
↓ 5 callersFunctionobserve_call
( observers: Option<ObservationSender>, execute: impl Future<Output = Result<CallOutput<R, H, C, E>, E
litellm-rust/crates/host/src/lifecycle.rs:149
↓ 5 callersFunctionobserve_unary
( observers: Option<ObservationSender>, execute: impl Future<Output = Result<R, E>>, )
litellm-rust/crates/host/src/lifecycle.rs:195
↓ 5 callersFunctionocr_binding
(native: NativeOcr | None)
tests/test_litellm_rust/ocr/test_dispatch.py:131
↓ 5 callersMethodof
(cls, config: RouterConfig, dynamic: RouterDynamicConfig)
litellm/router_rust/core.py:692
↓ 5 callersMethodof
(self, call_type: str)
tests/unit/responses/mcp/test_aresponses_api_with_mcp.py:2147
↓ 5 callersMethodof_rejected
(detail: str)
litellm/proxy/mcp_server/outbound_credentials/sso_assertion_refresher.py:213
↓ 5 callersFunctionofficial_client_outcomes
List then call through the official MCP client session, returning both outcomes.
tests/integration/_support/mcp.py:767
↓ 5 callersMethodon
(e,t)
litellm/proxy/_experimental/out/_next/static/chunks/2hk8n3azev--d.js:11
↓ 5 callersFunctionone_async_hook
(records: HookRecords, surface: str, outcome: str)
tests/integration/_support/hook_counts.py:317
↓ 5 callersMethodopenNode
(s)
litellm/proxy/swagger/swagger-ui-bundle.js:2
↓ 5 callersFunctionopenssl_enforces_fips
Python's OpenSSL runs with FIPS-only defaults and itself refuses scrypt.
litellm/proxy/common_utils/fips.py:27
↓ 5 callersMethodoperator_open_server_ids
Servers reachable through OPEN channels rather than a grant: operator-opened ``allow_all_keys`` servers, plus the caller's own active BYOM sub
litellm/proxy/mcp_server/mcp_server_manager.py:3632
↓ 5 callersFunctionoptional_str
(value: object)
litellm/proxy/pass_through_endpoints/llm_provider_handlers/batch_attribution.py:41
↓ 5 callersFunctionorgDetailHref
(orgId: string)
ui/litellm-dashboard/src/utils/entityLinks.ts:24
↓ 5 callersFunctionorganization_max_budget_check
Check if the organization is over its max budget. This function checks the organization budget using: 1. First, tries to use valid_token
litellm/proxy/auth/auth_checks.py:7560
↓ 5 callersFunctionorganization_member_add
[BETA] Add new members (either via user_email or user_id) to an organization If user doesn't exist, new user row will also be added to
litellm/proxy/management_endpoints/organization_endpoints.py:1315
↓ 5 callersFunctionoutbound_request
Header credentials are already in `headers`; SigV4 is applied here, over the bytes that are sent.
litellm-rust/crates/inference/src/outbound.rs:24
↓ 5 callersFunctionoutcome
()
tests/unit/proxy/test_route_llm_request.py:2637
↓ 5 callersMethodoutcome_callbacks
Router-core's ``OutcomeCallbacks`` for ``request``, empty (its default) while the outcome callbacks are listed. Past that only the request's b
litellm/router_rust/steps/callbacks.py:408
↓ 5 callersFunctionover
(tokens: u64)
litellm-rust/crates/cost/src/price.rs:296
↓ 5 callersFunctionoverlay_served_output_texts
( response_obj: dict[str, object] | str | list[object] | None, served_texts: object )
litellm/litellm_core_utils/served_output_texts.py:172
↓ 5 callersFunctionowned_connect_sink
(directory: Path)
tests/integration/_support/otlp_sink.py:516
↓ 5 callersFunctionowner
(**_: object)
tests/unit/proxy/proxy_server/test_routes_model_info.py:1715
↓ 5 callersFunctionp
(e,r,a,n,s,l)
litellm/proxy/_experimental/out/_next/static/chunks/218jq7ygtsrfo.js:1
↓ 5 callersFunctionp
(e)
litellm/proxy/_experimental/out/_next/static/chunks/0dvgjd47fte1d.js:1
↓ 5 callersMethodparse
(value: &Value)
litellm-rust/crates/router-core/src/access.rs:60
↓ 5 callersMethodparse
The raw token, as a caller passes it in `api_key`.
litellm-rust/crates/llms/src/anthropic/common_utils.rs:67
↓ 5 callersMethodparseKeyResponse
(resp map[string]interface{})
terraform/provider/litellm/client.go:258
↓ 5 callersFunctionparse_body
(response: httpx.Response)
litellm/llms/vertex_ai/gemini/long_decode.py:333
↓ 5 callersFunctionparse_canonical_date_range
(start_date: str | None, end_date: str | None)
litellm/proxy/management_endpoints/common_daily_activity.py:101
↓ 5 callersFunctionparse_content_for_reasoning
Parse the content for reasoning Returns: - reasoning_content: The content of the reasoning - content: The content of the message
litellm/litellm_core_utils/prompt_templates/reasoning_utils.py:344
↓ 5 callersFunctionparse_control_options
(kwargs: Mapping[str, object])
litellm/litellm_core_utils/get_litellm_params.py:133
↓ 5 callersFunctionparse_control_plane_replica_urls
The replicas a management read-back polls. LITELLM_CONTROL_PLANE_REPLICA_URLS names them outright; unset, they follow the two base URLs: every dat
tests/e2e/e2e_config.py:48
↓ 5 callersFunctionparse_session_update
(payload: str, expected_model: str, *, drop_params: bool = False)
litellm/llms/meta/realtime/transformation.py:234
↓ 5 callersFunctionparse_str_bool
(value: &str)
litellm-rust/crates/core-utils/src/serde_compat.rs:7
↓ 5 callersFunctionparse_tools
(json: &str)
litellm-rust/crates/token-counter/src/tools.rs:179
↓ 5 callersFunctionparse_unified_id
Parse a unified ID into its components. Args: unified_id: The unified ID (encoded or decoded) Returns: Dictionary with
litellm/llms/base_llm/managed_resources/utils.py:357
↓ 5 callersFunctionparse_upstream_reported_model
Read the model name an upstream reported, or ``None`` when it sent none. An empty or whitespace-only value counts as absent. A repeated header re
litellm/proxy/pass_through_endpoints/upstream_usage_headers.py:142
↓ 5 callersFunctionpart
(wire: Value)
litellm-rust/crates/llms-types/src/formats/messages/content.rs:769
↓ 5 callersMethodpath
(&self, name: &str)
litellm-rust/crates/http/tests/websocket.rs:71
↓ 5 callersFunctionpath_of
(endpoint: Endpoint)
tests/integration/providers/_gemini_long_decode_support.py:222
↓ 5 callersFunctionpathed
(op: str, path: str, value: JsonValue | None = None)
tests/integration/management/test_scim_group_pathless_patch.py:42
↓ 5 callersMethodpause
(self)
litellm/proxy/shutdown/scheduled_jobs.py:22
↓ 5 callersMethodpause
(e,t)
litellm/proxy/_experimental/out/_next/static/chunks/2hk8n3azev--d.js:14
↓ 5 callersFunctionpayload
(value: object)
tests/test_litellm_rust/cache/support.py:24
↓ 5 callersFunctionpct
(value: number)
ui/litellm-dashboard/src/app/(dashboard)/cost-optimization/_components/ShadowEvalSection.tsx:28
↓ 5 callersFunctionpersonal_key_owner_teams_with_status
The owner's current, loadable, unblocked, within-budget teams: the grant sources a personal key inherits MCP server, MCP tool and agent access fro
litellm/proxy/auth/auth_checks.py:6241
↓ 5 callersFunctionpersonal_key_team_model_access_applies
(valid_token: UserAPIKeyAuth | None)
litellm/proxy/auth/auth_checks.py:5966
↓ 5 callersMethodpersonal_key_team_server_grants
( auth: UserAPIKeyAuth, )
litellm/proxy/mcp_server/auth/user_api_key_auth_mcp.py:2782
↓ 5 callersMethodpick_input
The ``messages`` and ``input`` the endpoint's attempt helper hands its pick, which every selection stage reads: the pre-call checks run when e
litellm/router_rust/invoke/endpoints.py:257
↓ 5 callersFunctionpicked_id
The id of the deployment a selector returned.
litellm/router_rust/steps/selection.py:367
↓ 5 callersMethodping
(self)
litellm/caching/caching.py:1046
↓ 5 callersFunctionpinned_billing_time
Price every rate lookup inside this block at ``moment`` rather than at each one's own clock read.
litellm/_internal_context.py:211
↓ 5 callersFunctionpipeline_projection
(events: Sequence[FunctionTraceEvent])
tests/rust-python-harness/shared/tracing/steps.py:17
↓ 5 callersFunctionplain_request
(path: str)
tests/e2e_harness/test_fixture_bundle.py:56
↓ 5 callersFunctionpointer_child
(parent: &str, key: &str)
litellm-rust/crates/model-catalog/src/error.rs:254
↓ 5 callersMethodpolicy
(self, trigger: BaseException | None = None)
litellm/router_rust/invoke/streaming.py:143
↓ 5 callersFunctionpoll_until_blocked_stream
poll_until_blocked for raw/streamed sends, which return a StreamingResponse instead of a Result: retry while the call still succeeds (the data-pla
tests/e2e/guardrails/guardrails_client.py:644
↓ 5 callersFunctionpoll_until_guardrail_applied
( call: Callable[[], StreamingResponse], guardrail_name: str, *, timeout: float = POLL_TIMEOUT
tests/e2e/guardrails/guardrails_client.py:592
↓ 5 callersMethodpop
Drains the Redis list; a cancellation mid-drain puts the popped entries back before propagating.
litellm/proxy/db/request_error_tracking.py:266
↓ 5 callersFunctionpop_aws_auth_params
( optional_params: MutableMapping[str, object], # mutable-ok: pops the aws_* keys out of the caller's map
litellm/llms/bedrock/base_aws_llm.py:113
↓ 5 callersFunctionpop_vertex_request_labels
Resolve labels from optional ``labels`` (Gemini-style) and/or ``litellm_params["metadata"]`` / ``litellm_params["litellm_metadata"]`` (``
litellm/llms/vertex_ai/common_utils.py:227
↓ 5 callersFunctionpost_call
( logger: LoggingSurface, original_response: str, api_key: str | None, additional_args: dict[str, object]
litellm/rust_bridge/host/callbacks_legacy_python.py:241
↓ 5 callersFunctionpre_call
()
tests/unit/proxy/hooks/test_parallel_request_limiter_v3.py:4362
↓ 5 callersMethodpre_call
(self, input, api_key, model=None, additional_args={})
litellm/litellm_core_utils/litellm_logging.py:1679
↓ 5 callersMethodpre_call_check
(self, deployment: dict)
tests/unit/test_router/core/test_gaps_l2800.py:382
↓ 5 callersMethodprefix
(&self)
litellm-rust/crates/gateway-mcp/src/native/registry.rs:14
↓ 5 callersMethodprefix_positions
One cache key per content block of the cacheable prefix, oldest block first. Each key hashes the prefix content up to and including
litellm/router_utils/prompt_caching_cache.py:332
↓ 5 callersFunctionprepare
( identity: &AuthenticatedRequest, body: Map<String, Value>, )
litellm-rust/crates/gateway-inference/src/caching.rs:24
↓ 5 callersMethodprepare
The keyword view the rest of the call reads: a copy, so the deployment hook's own dict is left as the hook returned it, carrying the logger as `@clien
litellm-rust/crates/callbacks-legacy-python/src/adapter.rs:211
↓ 5 callersFunctionprepare_response_cost
(logging_obj: Logging, response: object, now: datetime)
litellm/proxy/spend_tracking/cost_finalization.py:14
← previousnext →16,301–16,400 of 202,479, ranked by callers