MCPcopy Create free account

hub / github.com/BerriAI/litellm / functions

Functions202,479 in github.com/BerriAI/litellm

↓ 7 callersFunctiondecrypt_credentials
Decrypt all secret fields in an MCPCredentials dict using the global salt key.
litellm/proxy/mcp_server/db.py:824
↓ 7 callersFunctiondefers_batch_cost
A batch retrieve before the batch's cost is final charges nothing, guardrails included, since the retrieve that settles the batch charges its whol
litellm/litellm_core_utils/batch_utils.py:384
↓ 7 callersFunctiondelete_fallback
Delete fallback configuration for a specific model. **Parameters:** - `model`: The model name to delete fallbacks for - `fallback_ty
litellm/proxy/management_endpoints/fallback_management_endpoints.py:388
↓ 7 callersFunctiondelete_model
( model_info: ModelInfoDelete, user_api_key_dict: UserAPIKeyAuth = Depends(user_api_key_auth), )
litellm/proxy/management_endpoints/model_management_endpoints.py:2940
↓ 7 callersMethoddelete_org
(self, organization_id: str)
tests/e2e/management/management_client.py:615
↓ 7 callersMethoddeliverable
The subset of ``destinations`` this fan-out can actually export to. A destination whose exporter will not build (a protocol whose package is
litellm/integrations/otel/plumbing/providers.py:990
↓ 7 callersFunctiondelta
(identity: str, text: str)
tests/integration/_support/responses_stream.py:69
↓ 7 callersMethoddeny
(self, reason: str = "")
litellm/harness/types.py:114
↓ 7 callersFunctiondeployment
(value: T)
litellm-rust/crates/auth-azure/src/native.rs:631
↓ 7 callersFunctiondeployment_reasoning_effort_levels
Only the deployment's own declaration counts. The router copies it onto the shared provider/model map key, and map entries describe a model rather
litellm/litellm_core_utils/reasoning_effort_capability.py:128
↓ 7 callersMethoddeployments
(self, name: str)
tests/integration/spend/test_router_cooldown_failure_row.py:65
↓ 7 callersFunctiondereference
(s,o)
litellm/proxy/swagger/swagger-ui-bundle.js:2
↓ 7 callersMethodderegister_agent
(self, agent_name: str)
litellm/proxy/agent_endpoints/agent_registry.py:436
↓ 7 callersMethoddestroy
()
litellm/proxy/_experimental/out/_next/static/chunks/3lb5gjzjxjxem.js:1
↓ 7 callersMethoddisable_spend_updates
returns True if should not update spend in db Skips writing spend logs and updates to key, team, user spend to DB
litellm/proxy/utils.py:8829
↓ 7 callersMethoddispatch_attempt_failure_handlers
( self, exception: Exception, traceback_exception: str )
litellm/litellm_core_utils/litellm_logging.py:2756
↓ 7 callersMethoddispatch_success_handlers
Route success logging to async and/or sync handlers for this request. ``prefer_async_handlers`` only bypasses the sync-SDK-only shortcut (e.g
litellm/litellm_core_utils/litellm_logging.py:2554
↓ 7 callersMethoddispatch_success_handlers
(self, result: list[dict], prefer_async_handlers: bool = False)
tests/unit/litellm_core_utils/test_realtime_streaming.py:4839
↓ 7 callersMethoddoes_team_member_have_permissions_for_endpoint
Raises an exception if the team member does not have permissions for calling the endpoint for a team
litellm/proxy/management_helpers/team_member_permission_checks.py:112
↓ 7 callersMethoddone
()
litellm/proxy/_experimental/out/_next/static/chunks/085r5ieqhajig.js:1
↓ 7 callersFunctiondownloadBlob
(blob: Blob, filename: string)
ui/litellm-dashboard/src/components/EntityUsageExport/utils.ts:16
↓ 7 callersMethoddownload_file
(self, storage_url: str)
tests/unit/proxy/batches_endpoints/test_litellm_executed_batches.py:273
↓ 7 callersFunctiondrop_non_python_regex_patterns
Drop every regex in a schema position that Python's ``re`` cannot compile. OpenAI validates tool ``parameters`` against the 2020-12 metaschema wi
litellm/litellm_core_utils/prompt_templates/json_schema_utils.py:189
↓ 7 callersFunctiondrop_tool_reference_parts_from_tool_messages
Remove tool_reference content parts from role:"tool" messages. The OpenAI chat spec only accepts text in tool messages, so a tool_reference
litellm/litellm_core_utils/prompt_templates/common_utils.py:1488
↓ 7 callersFunctiondrops_params
(litellm_params: GenericLiteLLMParams)
litellm/llms/openai/responses/transformation.py:128
↓ 7 callersFunctiondumps
`json.dumps(value)`.
litellm-rust/crates/python-compat/src/json.rs:19
↓ 7 callersMethodduplicated
(self)
tests/integration/observability/_azure_storage_support.py:165
↓ 7 callersFunctionduration
A TTL in seconds, raising what `int(seconds)` raises for infinity and NaN.
litellm-rust/crates/python-bridge/src/redis/store/mod.rs:80
↓ 7 callersFunctiondynamic_otlp_endpoint
Per-request OTLP endpoint for ``callback_name``, or ``None`` if N/A. ``None`` means "keep the preset's own endpoint".
litellm/integrations/otel/presets/__init__.py:134
↓ 7 callersFunctione3
(e)
litellm/proxy/_experimental/out/_next/static/chunks/0ayyip3c-drm1.js:1
↓ 7 callersFunctione9
(e)
litellm/proxy/_experimental/out/_next/static/chunks/0_6jqgnorqxl2.js:1
↓ 7 callersFunctione_
(e)
litellm/proxy/_experimental/out/_next/static/chunks/0ayyip3c-drm1.js:1
↓ 7 callersFunctionea
litellm/proxy/_experimental/out/_next/static/chunks/3yrgf44a9g52e.js:1
↓ 7 callersFunctionecho
()
litellm-rust/crates/gateway-forward/tests/websocket.rs:63
↓ 7 callersFunctionee
litellm/proxy/_experimental/out/_next/static/chunks/11et800l-khs-.js:1
↓ 7 callersMethodeffective_max_budget
(self, now: datetime)
litellm/models/budget.py:49
↓ 7 callersFunctioneh
(e)
litellm/proxy/_experimental/out/_next/static/chunks/0ayyip3c-drm1.js:1
↓ 7 callersFunctionei
(e,t)
litellm/proxy/_experimental/out/_next/static/chunks/2_9r2jeem0dh8.js:1
↓ 7 callersFunctionel
(e,t)
litellm/proxy/_experimental/out/_next/static/chunks/11et800l-khs-.js:1
↓ 7 callersFunctioneligible_deployments
( model: str, deployments: Sequence[_Deployment] )
litellm/router_utils/endpoint_eligibility.py:111
↓ 7 callersFunctionembeddings
(response: httpx.Response)
tests/integration/providers/test_gemini_embedding_file_block_wire.py:86
↓ 7 callersFunctionembeds_thought_signature_in_tool_call_id
(litellm_params: Mapping[str, object])
litellm/litellm_core_utils/model_param_helper.py:49
↓ 7 callersFunctionemit_cassette_cache_session_banner
Call from ``pytest_terminal_summary``. No-op on xdist workers.
tests/_vcr_conftest_common.py:1946
↓ 7 callersFunctionen
({accessToken:e,teams:t,possibleUIRoles:r,onUsersCreated:i})
litellm/proxy/_experimental/out/_next/static/chunks/37cr8te-omsl0.js:1
↓ 7 callersFunctionencode_all
(mut codec: C, items: impl IntoIterator<Item = I>)
litellm-rust/crates/framer/tests/support/mod.rs:9
↓ 7 callersMethodencode_bedrock_runtime_modelid_arn
Encodes any "/" found in the modelId of an AWS Bedrock Runtime Endpoint when arns are passed in. - modelID value can be an ARN which
litellm/passthrough/utils.py:108
↓ 7 callersFunctionencode_character_id_in_response
(response: Any, custom_llm_provider: str, model_id: str | None)
litellm/proxy/video_endpoints/utils.py:66
↓ 7 callersMethodencode_model_id
Double encode the model ID to ensure it matches the expected double-encoded format. Args: model_id (str): The model ID to
litellm/llms/bedrock/base_aws_llm.py:635
↓ 7 callersMethodencode_queries
(self, docs: list[str], **kwargs)
litellm/router_strategy/auto_router/litellm_encoder.py:114
↓ 7 callersFunctionencode_ui_session_jwt
Encode a UI session cookie JWT with a bounded ``exp``. The single choke point every UI login path (SSO and username/password /login, /v2, /v3
litellm/proxy/auth/login_utils.py:503
↓ 7 callersFunctionenforce_output_token_estimates_are_admin_only
Only a proxy admin may change what a key or team declares its models emit. That declaration is what the TPM limiter reserves for a request omitti
litellm/proxy/auth/auth_utils.py:1587
↓ 7 callersMethodenforce_provider_file_access
Deny access to a provider-format file id owned by another caller. Ownership rows for provider-format ids are written when a managed b
enterprise/litellm_enterprise/proxy/hooks/managed_files.py:1101
↓ 7 callersFunctionensure_mappers
Return ``mapper_names`` with each of ``names`` appended if not already present. Order is preserved and duplicates are skipped, so composing sever
litellm/integrations/otel/presets/utils.py:9
↓ 7 callersMethodenum
()
litellm/proxy/_experimental/out/_next/static/chunks/2j7hjtph8n74b.js:1
↓ 7 callersMethoderror
(&self, py: Python<'_>, error: Error)
litellm-rust/crates/python-bridge/src/routes/inference.rs:119
↓ 7 callersMethoderror
(&self)
litellm-rust/crates/python-compat/src/json.rs:80
↓ 7 callersFunctionerror_class
Textract answers both an unsupported format and a multi-page PDF or TIFF with a bare "unsupported document format", which reads like a corrupt file. S
litellm-rust/crates/llms/src/aws_textract/ocr/common_utils.rs:325
↓ 7 callersFunctionerror_response_text
(response: httpx.Response)
litellm/llms/bedrock/common_utils.py:89
↓ 7 callersFunctioneu
(e,t,r)
litellm/proxy/_experimental/out/_next/static/chunks/2_9r2jeem0dh8.js:1
↓ 7 callersFunctionevent_stream_reply
(payloads: tuple[bytes, ...], *, then: tuple[bytes, ...] = ())
tests/integration/providers/_sagemaker_chat_container.py:109
↓ 7 callersFunctionevents
(response_id: str, text: str)
tests/integration/_support/responses_ws_peer.py:45
↓ 7 callersFunctionevict_local
Drop cached objects from this worker's memory alone, before any Redis round trip, so a deleted key stops authenticating on the handling worke
litellm/proxy/common_utils/auth_cache_invalidation_pubsub.py:182
↓ 7 callersMethodexec
( self, prompt: str, *, model: str, cwd: Path, compact_after_t
tests/e2e/codex/harness.py:74
↓ 7 callersMethodexecute
( &self, request: ChatCompletionsRequest<'_>, interceptors: &impl litellm_host::interc
litellm-rust/crates/inference-chat/src/lib.rs:45
↓ 7 callersFunctionexecute_statement
( client: &Client, connection: &Connection, sql: &str, timeout: Duration, )
litellm-rust/crates/storage-clickhouse/src/migrate.rs:17
↓ 7 callersFunctionextract_and_raise_litellm_exception
Covers scenario where litellm sdk calling proxy. Enables raising the special errors raised by litellm, eg. ContextWindowExceededError.
litellm/litellm_core_utils/exception_mapping_utils.py:239
↓ 7 callersMethodextract_client_ip_from_xff_hops
Resolve the originating client IP from an X-Forwarded-For chain by counting ``num_trusted_hops`` entries from the right. Eac
litellm/proxy/auth/ip_address_utils.py:230
↓ 7 callersFunctionextract_file_parts_from_images
( images: Sequence[object] | None, size_limit: int | None = None )
litellm/proxy/guardrails/guardrail_hooks/ovalix/ovalix_extraction.py:203
↓ 7 callersFunctionextract_text_from_a2a_response
Extract text content from A2A response result. Args: response_dict: A2A response dict with 'result' containing message max_d
litellm/llms/a2a/common_utils.py:100
↓ 7 callersFunctionfailed
(identity: str, code: str, message: str)
tests/integration/_support/responses_stream.py:61
↓ 7 callersFunctionfailure_spend
(gateway: Gateway, key: str, rows: int = 1)
tests/integration/spend/_partial_stream.py:150
↓ 7 callersFunctionfallback_lookup_groups
Ordered keys for resolving a fallback chain: the tier a pre-routing hook selected wins, then the routed group, then the requested group, then
litellm/router_utils/fallback_event_handlers.py:568
↓ 7 callersFunctionfetch_latest_health_checks
(prisma_client: PrismaClient)
litellm/proxy/db/health_check_latest.py:88
↓ 7 callersFunctionfetch_model_ids
( base_url: str, api_key: str, *, get: Callable[..., requests.Response] = requests.get, he
litellm/proxy/client/cli/commands/pi.py:126
↓ 7 callersFunctionfetch_model_limits
Best effort: pi, OpenCode and Codex fall back to their own defaults for models without limits or reasoning efforts, so an unavailable /model_group
litellm/proxy/client/cli/commands/pi.py:140
↓ 7 callersFunctionfill_missing_api_key_aliases
Fill null api_key_alias / team_id / user_email on export rows whose api_key was double-hashed. Used by CloudZero and Focus, which join D
litellm/proxy/spend_tracking/key_metadata_recovery.py:530
↓ 7 callersMethodfill_reasoning_content
Moonshot reasoning models require `reasoning_content` on every assistant message that contains tool_calls (multi-turn tool-calling fl
litellm/llms/moonshot/chat/transformation.py:162
↓ 7 callersFunctionfilter_request_log_index_diff
The `prisma migrate diff` script without the statements that create a migration-job-owned index, which the schema declares and the migrations deli
litellm-proxy-extras/litellm_proxy_extras/request_log_indexes.py:131
↓ 7 callersFunctionfilter_reserved_deployments
( healthy_deployments: Sequence[_DeploymentT], request_team_id: str | None, now: datetime, )
litellm/litellm_core_utils/access_windows.py:63
↓ 7 callersMethodfilter_server_ids_by_ip_with_info
Filter server IDs by client IP — external callers only see public servers. Returns (filtered_ids, ip_blocked_count) where ip_blocked
litellm/proxy/mcp_server/mcp_server_manager.py:3967
↓ 7 callersMethodfinalize_streaming_generator_cleanup
( request: Request | None, request_data: dict, response: object, stream_comple
litellm/proxy/common_request_processing.py:4820
↓ 7 callersMethodfindWrapping
(e)
litellm/proxy/_experimental/out/_next/static/chunks/085r5ieqhajig.js:1
↓ 7 callersMethodfind_many
( self, where: Mapping[str, object] | None = None, order: Mapping[str, str] | None = N
litellm/proxy/agent_endpoints/agent_registry.py:112
↓ 7 callersFunctionfired_exactly
The events as written: ``5 == 5.0``, and the cooldown's type is part of the event.
tests/unit/test_router/core/test_parity_cooldown_service_model.py:138
↓ 7 callersFunctionfirst_item
(value: JsonValue)
tests/integration/providers/_api_base_trailing_slash_support.py:300
↓ 7 callersFunctionflag
A flag every Python surface reads the same way. `_str_to_bool` takes only `true`, `1` and `yes` as true while the proxy's pydantic models also take `o
litellm-rust/crates/redis-resolve/src/options.rs:208
↓ 7 callersFunctionflatten_unencrypted_web_search_results_in_anthropic_messages
Return a new message list with replayed ``web_search_tool_result`` blocks that carry no ``encrypted_content`` rewritten into plain ``text`` b
litellm/llms/anthropic/common_utils.py:2657
↓ 7 callersMethodflush
(self)
tests/unit/proxy/hooks/test_model_max_budget_limiter.py:445
↓ 7 callersFunctionflush_langfuse_tracing
Force-flush every export channel this process acquired, all within one ``timeout_millis`` deadline. ``True`` only when every channel flushed in t
litellm/integrations/langfuse/langfuse_sdk.py:990
↓ 7 callersMethodflush_session_to_db
Drain session aggregator and upsert into LiteLLM_AdaptiveRouterSession. Returns number of session rows flushed.
litellm/router_strategy/adaptive_router/update_queue.py:151
↓ 7 callersMethodfor_model
This config when ``model`` is served on the OpenAI Responses surface, else ``None``. The capability decision lives here rather than in the sh
litellm/llms/bedrock/responses/transformation.py:199
↓ 7 callersMethodforced_tool_use_unsupported
(model: str)
litellm/llms/anthropic/common_utils.py:965
↓ 7 callersMethodforward
( self, mount: str, method: str, url: str, headers: dict[str, str], body: bytes | None, timeout: float
tests/e2e/provider_cache.py:568
↓ 7 callersMethodfromReplace
(e,t,n,s)
litellm/proxy/_experimental/out/_next/static/chunks/085r5ieqhajig.js:1
↓ 7 callersMethodft
(self, index_name)
tests/unit/llms/valkey/vector_stores/test_valkey_transformation.py:35
↓ 7 callersMethodfully_unhealthy_model_names
Groups whose every deployment failed its health check (was ``async_get_fully_unhealthy_model_names``).
litellm/router_api.py:1348
↓ 7 callersFunctionfunction_call_item
( identity: str, arguments: JsonValue = ARGUMENTS, name: str = "get_weather" )
tests/integration/streaming/faked_responses_case.py:83
↓ 7 callersFunctiongateway_endpoint
(recorder: Recorder, **kwargs: Any)
tests/unit/harness/test_endpoint.py:70
↓ 7 callersMethodgetBlob
(path: string, options?: RequestOptions)
ui/litellm-dashboard/src/lib/http/client.ts:128
↓ 7 callersFunctiongetCookie
(name: string)
ui/litellm-dashboard/src/utils/cookieUtils.ts:134
← previousnext →10,801–10,900 of 202,479, ranked by callers