MCPcopy Create free account

hub / github.com/BerriAI/litellm / functions

Functions202,479 in github.com/BerriAI/litellm

↓ 13 callersFunctionget_team_models
Returns: - List of model name strings - Empty list if no models set - If model_access_groups is provided, only return models that are
litellm/proxy/auth/model_checks.py:189
↓ 13 callersFunctionget_thought_signature_from_tool
Extract thought signature from tool call's provider_specific_fields. If not provided try to extract thought signature from tool call id Chec
litellm/litellm_core_utils/prompt_templates/factory.py:1274
↓ 13 callersMethodget_tool
Get a tool from the registry by name
litellm/proxy/mcp_server/tool_registry.py:76
↓ 13 callersMethodget_usage_as_dict
Like get_usage_from_response_obj but returns a plain dict, skipping the Pydantic Usage construction on the hot path.
litellm/litellm_core_utils/litellm_logging.py:7168
↓ 13 callersMethodget_vertex_region
(vertex_region: str | None, model: str)
litellm/llms/vertex_ai/vertex_llm_base.py:138
↓ 13 callersFunctiongithub_host_endpoints
(host: str)
litellm/llms/github_copilot/github_host.py:47
↓ 13 callersFunctionhandshake
(target: &str, headers: &str)
litellm-rust/crates/gateway-forward/tests/websocket.rs:99
↓ 13 callersFunctionhas_headroom_retrieve_tool
(tools: object)
litellm/proxy/guardrails/guardrail_hooks/headroom/headroom.py:387
↓ 13 callersMethodhealth
(&self)
litellm-rust/crates/router-core/src/lib.rs:418
↓ 13 callersFunctionhoist_additional_tools
( input: str | ResponseInputParam, tools: Sequence[ALL_RESPONSES_API_TOOL_PARAMS] | None, *, d
litellm/responses/additional_tools.py:96
↓ 13 callersFunctioni
litellm/proxy/_experimental/out/_next/static/chunks/2k-fkubkt19cc.js:8
↓ 13 callersMethodingest
Execute the full ingestion pipeline. Args: file_data: Tuple of (filename, content_bytes, content_type) file_
litellm/rag/ingestion/base_ingestion.py:299
↓ 13 callersMethodinitialize_scheduled_background_jobs
Initializes scheduled background jobs
litellm/proxy/proxy_server.py:11604
↓ 13 callersMethodinsert
(self, index: SupportsIndex, value: object)
litellm/router_rust/surface/indexes.py:108
↓ 13 callersMethodinto_string
(self)
litellm-rust/crates/core-utils/src/url_utils.rs:67
↓ 13 callersFunctionisForecastClassifier
(type: ClassifierType)
ui/litellm-dashboard/src/components/add_model/forecast_classifier_config.ts:82
↓ 13 callersFunctionisKeyed
(s)
litellm/proxy/swagger/swagger-ui-bundle.js:2
↓ 13 callersMethodis_async_callable
Fast, cached check for whether a callback is an async function. Falls back gracefully if the object cannot be weak-referenced or cached.
litellm/litellm_core_utils/coroutine_checker.py:22
↓ 13 callersMethodis_sensitive_key
(self, key: str, excluded_keys: set[str] | None = None)
litellm/litellm_core_utils/sensitive_data_masker.py:80
↓ 13 callersFunctionis_token_within_budget_for_model
True when the key and the user behind it can still pay for `model`. A zero-cost fallback target is always allowed: refusing it would deny a
litellm/proxy/auth/fallback_budget.py:121
↓ 13 callersFunctionis_user_key_cache_key
Only user-key objects are cached under a bare ``hash_token`` digest; every other object uses a prefixed key.
litellm/proxy/common_utils/user_api_key_cache.py:37
↓ 13 callersFunctionitems_of
(value: JsonValue)
tests/integration/providers/_gemini_long_decode_support.py:289
↓ 13 callersFunctionjob_to_charge_once
A finished batch or video reports its whole cost on every poll, so it is charged once per counter.
litellm/litellm_core_utils/batch_utils.py:434
↓ 13 callersMethodjson
(self)
litellm/integrations/gitlab/gitlab_client.py:46
↓ 13 callersFunctionkey_model_aliases_for_auth_check
(valid_token: UserAPIKeyAuth | None)
litellm/proxy/auth/auth_checks.py:5291
↓ 13 callersFunctionkey_window
(key: str, model: str, period: str)
tests/integration/management/test_budget_reset_windows.py:156
↓ 13 callersFunctionl
litellm/proxy/_experimental/out/_next/static/chunks/11et800l-khs-.js:1
↓ 13 callersMethodlabeled_tiers
Every tier paired with its display name, in ascending severity order.
litellm/router_strategy/complexity_router/config.py:2366
↓ 13 callersMethodlanded
(self, response_ids: Sequence[str])
tests/integration/observability/test_signoz_delivery.py:235
↓ 13 callersFunctionload_catalog
(root: Path | None = None)
tests/rust-python-harness/cli/catalog.py:97
↓ 13 callersMethodlog_event_on_langfuse
( self, kwargs, response_obj, start_time=None, end
tests/unit/integrations/test_langfuse.py:90
↓ 13 callersFunctionlookup
( cache: DualCache, scope: str, prefix: PromptPrefix, now: float | None = None )
litellm/llms/anthropic/prompt_cache_prediction.py:802
↓ 13 callersFunctionmake_closer
(clock: FakeClock, grace_seconds: float = 60.0)
tests/unit/caching/test_evicted_client_closer.py:83
↓ 13 callersFunctionmake_ctx
(tmp_path: Path, **kwargs: Any)
tests/unit/harness/handlers/test_deepagents_handler.py:67
↓ 13 callersFunctionmanaged_agent_policy
The admitted managed policy, or ``None`` when the subject was never admitted as a managed agent. ``admit_managed_actor`` only assigns ``managed_a
litellm/proxy/agent_endpoints/auth/managed_authorization.py:149
↓ 13 callersMethodmark_logging_complete
( self, event_type: Literal["async_success", "sync_success", "async_failure", "sync_failure"],
litellm/litellm_core_utils/litellm_logging.py:2946
↓ 13 callersFunctionmarker
Any strategy-router marker: ``model`` is its ``auto_router/...`` litellm model.
tests/unit/test_router/core/gaps_l11100_helpers.py:74
↓ 13 callersFunctionmarker_of
(request: Request)
tests/integration/_support/bedrock_runtime_peer.py:40
↓ 13 callersFunctionmarkers_received
(wire: Wire)
tests/integration/routing/test_payment_required_cooldown.py:94
↓ 13 callersMethodmatchType
(e)
litellm/proxy/_experimental/out/_next/static/chunks/085r5ieqhajig.js:1
↓ 13 callersMethodmessages_with_default_injections
Return the messages auto prompt caching will send, default breakpoints included. Router cache affinity depends on this. Deployment selection
litellm/integrations/anthropic_cache_control_hook.py:1620
↓ 13 callersFunctionmint_envelope
Seal ``grant`` for ``identity`` into a client-held envelope. ``exp`` is ``grant.expires_in`` seconds from ``now`` when the upstream reports a
litellm/proxy/mcp_server/outbound_credentials/envelope.py:311
↓ 13 callersFunctionminted
(result: object)
tests/unit/llms/base_llm/auth/test_shared_token_store.py:77
↓ 13 callersMethodmodel_dump
(self)
litellm/llms/base_llm/guardrail_translation/base_translation.py:21
↓ 13 callersMethodnamespace_tool_name_map
(tools: ResponseTools)
litellm/responses/litellm_completion_transformation/transformation.py:3082
↓ 13 callersFunctionnewTeamResourceData
(t *testing.T, raw map[string]interface{})
terraform/provider/litellm/resource_team_test.go:268
↓ 13 callersFunctionnew_marker
()
tests/integration/providers/_anthropic_safeguards_beta_support.py:88
↓ 13 callersFunctionnormalize_reasoning_effort_value
Lower a tier the deployment does not accept to the nearest one it does, leaving others alone. A deployment whose own reasoning_effort_levels name
litellm/llms/anthropic/messages/utils.py:192
↓ 13 callersFunctionnow
()
litellm/proxy/swagger/swagger-ui-bundle.js:2
↓ 13 callersFunctiono
(e,t,r)
litellm/proxy/_experimental/out/_next/static/chunks/2aatum0yo4cv8.js:1
↓ 13 callersFunctionone_classifier_call
(rig: Rig, request_nonce: str)
tests/integration/_support/safeguards_rig.py:414
↓ 13 callersFunctiononly_backend_call
(rig: Rig, request_nonce: str)
tests/integration/messages_endpoint/providers/databricks/test_databricks_native_messages_wire.py:212
↓ 13 callersMethodopen
(self)
litellm/proxy/db/log_db_metrics.py:64
↓ 13 callersFunctionoperation
( name: &str, execute: impl Future<Output = Result<R, E>>, )
litellm-rust/crates/inference/src/diagnostic.rs:54
↓ 13 callersFunctionoptional_sequence
(value: object)
litellm/rust_bridge/public_call.py:41
↓ 13 callersFunctionowned_database
(password: str)
tests/integration/database/test_v1_migration_error_logging.py:101
↓ 13 callersFunctionowned_upstream
( directory: Path, *, tls: bool = False, port: int | None = None, inherited: Mapping[str, str] = os.enviro
tests/integration/_support/process.py:1147
↓ 13 callersFunctionparseErrorMessage
(error: any)
ui/litellm-dashboard/src/components/shared/errorUtils.tsx:8
↓ 13 callersFunctionparse_prompt_tokens_details
(usage: Usage)
litellm/litellm_core_utils/llm_cost_calc/utils.py:1062
↓ 13 callersMethodparse_response
(self, payload: object, request: DecisionsIRRequest)
litellm/llms/openai/decisions/transformation.py:375
↓ 13 callersFunctionparse_stalls
Holds in log order, each carrying the collector passes logged since the previous hold: a pass runs inside the task asyncio then blames and is only
tests/e2e/perf/setup/event_loop.py:139
↓ 13 callersFunctionparse_tool_call_arguments
Parse tool call arguments from a JSON string. When the JSON is malformed (e.g. truncated by the model), this function attempts a lightwe
litellm/litellm_core_utils/prompt_templates/tool_call_json_repair.py:70
↓ 13 callersMethodpattern_to_regex
Convert a wildcard pattern to a regex pattern example: pattern: openai/* regex: openai/.* pattern: openai/f
litellm/litellm_core_utils/pattern_match_deployments.py:113
↓ 13 callersFunctionpdf_document
(pages: Sequence[Page], **fields: JsonValue)
tests/integration/_support/pdf_document.py:109
↓ 13 callersFunctionpeer
(delivered: int | None, gate: threading.Event | None = None)
tests/integration/spend/_partial_stream.py:71
↓ 13 callersFunctionpersist_sso_identity_assertion
( user_id: str, assertion: SSOIdentityAssertion, cache: SSOAssertionCache = _ASSERTION_CACHE )
litellm/proxy/mcp_server/outbound_credentials/sso_assertion_store.py:317
↓ 13 callersFunctionpins_since
(redis: Redis, before: frozenset[str])
tests/integration/routing/_prompt_cache_pin_support.py:97
↓ 13 callersMethodpipe
(e)
litellm/proxy/_experimental/out/_next/static/chunks/2j7hjtph8n74b.js:1
↓ 13 callersMethodpop_next_value
Attempt to decode one complete JSON value from the front of the buffer. On success, advances a cursor past that value (keeping any
litellm/litellm_core_utils/json_fragment_accumulator.py:55
↓ 13 callersFunctionpost
( gateway: Gateway, path: str, body: Mapping[str, JsonValue], *, key: str | None = None,
tests/integration/compatibility/_status_code_audit.py:177
↓ 13 callersMethodpost
( self, url: str, data: dict | str | bytes | None = None, json: dict | None =
litellm/llms/custom_httpx/http_handler.py:960
↓ 13 callersMethodprepare
( self, plan: Plan, request: Mapping[str, object], model: str )
litellm/router_rust/invoke/endpoints.py:1519
↓ 13 callersFunctionprice_transcription_job
Amazon Transcribe bills every second of the media file, silence included, and reports no duration itself, so the job is polled to completion
litellm/proxy/pass_through_endpoints/llm_provider_handlers/transcribe_passthrough_logging_handler.py:417
↓ 13 callersMethodprocess_output_response
Process output response by applying guardrails to text content. Args: response: LiteLLM ModelResponse object
litellm/llms/openai/chat/guardrail_translation/handler.py:437
↓ 13 callersFunctionprojectMountedValues
( registry: MountRegistry, getValues: UseFormGetValues<MountedFormValues>, )
ui/litellm-dashboard/src/components/common_components/MountedFormField.tsx:84
↓ 13 callersFunctionprovider
(model: &str, provider: &str)
litellm-rust/crates/inference/src/diagnostic.rs:44
↓ 13 callersFunctionpublish_auth_cache_invalidation
Best-effort broadcast so every worker drops its local in-memory copy of a mutated management object; without this, only the handling worker a
litellm/proxy/common_utils/auth_cache_invalidation_pubsub.py:129
↓ 13 callersFunctionpublish_global_otel_v2_provider
Select the single v2 logger and publish its provider as the OTel global. The proxy calls this once at startup, after callbacks are initialized, s
litellm/integrations/otel/logger.py:941
↓ 13 callersFunctionpure_ctx
(tmp_path: Path, **overrides: Any)
tests/unit/llms/claude_code/harness/test_transformation.py:197
↓ 13 callersFunctionraise_for_xai_status
(response: httpx.Response)
litellm/llms/xai/batches/transformation.py:56
↓ 13 callersMethodraw
( self, endpoint: Endpoint, marker: str, stream: bool, key: str | None = None, trace_id: str | None =
tests/integration/observability/test_otel_excluded_services_matrix.py:227
↓ 13 callersFunctionread
Reads and projects the secret manager settings group in one attached operation.
litellm-rust/crates/python-bridge/src/secrets/config.rs:114
↓ 13 callersFunctionread_field
Reads one field from a dict or from an object. LiteLLM's replies arrive as Pydantic models on some paths and as plain dicts on others, depend
litellm/proxy/guardrails/guardrail_hooks/llm_shield_proxy/payload.py:90
↓ 13 callersFunctionrealtime_error_event
(message: str, error_type: str)
litellm/litellm_core_utils/realtime_errors.py:26
↓ 13 callersMethodrecord
Run `record` when `write` happens for this request, else record nothing.
litellm-rust/crates/router-core/src/outcome.rs:62
↓ 13 callersFunctionrecord_db_io
()
litellm/proxy/db/log_db_metrics.py:83
↓ 13 callersMethodrecord_gateway_injection
Name the deployment whose payload the gateway, not the client, put breakpoints on. Spend accounting only asks whether litellm acted, so what
litellm/integrations/anthropic_cache_control_hook.py:1725
↓ 13 callersMethodregion
(&self)
litellm-rust/crates/cache-s3/src/cache.rs:82
↓ 13 callersMethodregister
Register any guardrail via POST /guardrails and return its id, once every replica can be expected to serve it. New built-ins register with
tests/e2e/guardrails/guardrails_client.py:240
↓ 13 callersMethodremote
(&self, result: Result<T, Error>)
litellm-rust/crates/cache/src/dual.rs:98
↓ 13 callersMethodremove
(e)
litellm/proxy/_experimental/out/_next/static/chunks/0tx19jeq-wskb.js:2
↓ 13 callersMethodremove_callback_from_all_lists
(self, obj: object, require_self: bool = False)
tests/unit/test_router/core/gaps_l11100_helpers.py:23
↓ 13 callersFunctionrender
(value: ComplexityRouterConfigValue, onChange = vi.fn())
ui/litellm-dashboard/src/components/add_model/HeuristicScoringConfig.test.tsx:22
↓ 13 callersMethodrepair_invalid_indexes
Rebuild LiteLLM indexes an interrupted CREATE INDEX CONCURRENTLY left INVALID (a migration deadlock between replicas is the usual cause; the
litellm-proxy-extras/litellm_proxy_extras/utils.py:879
↓ 13 callersMethodreplaceChild
(e,t)
litellm/proxy/_experimental/out/_next/static/chunks/085r5ieqhajig.js:1
↓ 13 callersFunctionrequest_with_retry
Bounded retry on statuses attributable to the PROVIDER, never the proxy. The system under test is the proxy, so the transport may only absorb
tests/e2e/e2e_http.py:318
↓ 13 callersFunctionrescript
(handle: ScenarioHandle, response: JsonResponse | SseResponse)
tests/integration/caching/response_cache_case.py:270
↓ 13 callersMethodreserve
(self, size: int)
litellm/proxy/lens/buffering.py:21
↓ 13 callersFunctionresolve_model_budget
Find the `model_max_budget` entry that governs `model`, or None. An entry named after the model wins, then one named after a model access gro
litellm/proxy/hooks/model_max_budget_limiter.py:244
↓ 13 callersFunctionresolve_openai_bearer_token
( api_key: str | None, api_base: str | None, litellm_params: Mapping[str, object] | None, )
litellm/llms/openai/workload_identity.py:113
← previousnext →5,001–5,100 of 202,479, ranked by callers