MCPcopy Create free account

hub / github.com/BerriAI/litellm / functions

Functions202,479 in github.com/BerriAI/litellm

↓ 6 callersMethodasync_should_run_responses_agentic_loop
Check if WebSearch interception is needed for the Responses API.
litellm/integrations/websearch_interception/handler.py:879
↓ 6 callersFunctionatext_completion
Implemented to handle async streaming for the text completion endpoint
litellm/main.py:7984
↓ 6 callersFunctionattempt
()
tests/e2e/management/test_management_e2e.py:101
↓ 6 callersFunctionauditCoverage
(calls []endpointCall, specPaths map[string]map[string]json.RawMessage, allowlist map[string]bool)
terraform/provider/tools/endpointaudit/coverage.go:78
↓ 6 callersFunctionauth_callback
Verify login
litellm/proxy/management_endpoints/ui_sso.py:2298
↓ 6 callersFunctionauthorize_mcp_server
( resolved: ResolvedMCPServer | None, user_api_key_dict: UserAPIKeyAuth, *, manager: MCPServer
litellm/proxy/mcp_server/server_resolution.py:93
↓ 6 callersFunctionauthorize_oauth_credential_request
(request: Request, server_id: str)
litellm/proxy/mcp_server/bridge_token_flow.py:450
↓ 6 callersFunctionautorouter_savings_for_request
Return net savings for established usage, or None when the estimate is unavailable.
litellm/proxy/spend_tracking/savings.py:471
↓ 6 callersFunctionawait_hit
(gateway: Gateway, wire: Wire, body: Mapping[str, JsonValue], key: str)
tests/integration/caching/test_embedding_file_block_cache.py:104
↓ 6 callersFunctionawait_with_timeout
Await ``pending`` on the event loop and give it up at the deadline, even if it swallows cancellation.
litellm/proxy/guardrails/guardrail_hooks/custom_code/bounded_execution.py:217
↓ 6 callersFunctionazure_ad_token_provider
Reads `azure_ad_token_provider`, ignoring the falsy and non-callable values litellm's public API has always accepted in its place.
litellm-rust/crates/python-bridge/src/credentials.rs:35
↓ 6 callersFunctionazure_texts
(azure: Wire)
tests/integration/observability/azure_dispatch_support.py:138
↓ 6 callersFunctionbackfill_null_oauth2_flows
Classify every ``auth_type=oauth2`` row whose ``oauth2_flow`` is null; stamp the provable ones, warn on the ambiguous ones, and return counts per
litellm/proxy/mcp_server/oauth2_flow_backfill.py:152
↓ 6 callersMethodbase_passthrough_process_llm_request
( self, request: Request, fastapi_response: Response, user_api_key_dict: UserA
litellm/proxy/common_request_processing.py:3842
↓ 6 callersFunctionbase_url
(gateway: Gateway)
tests/integration/routing/test_fallback_hop_foreign_encrypted_reasoning.py:440
↓ 6 callersMethodbatch_creation_handler
Handle Anthropic batch creation passthrough logging. Creates a managed object for cost tracking when batch job is successfully create
litellm/proxy/pass_through_endpoints/llm_provider_handlers/anthropic_passthrough_logging_handler.py:590
↓ 6 callersFunctionbatch_upload_form
(provider: str, *, target_model_names: str | None = None)
tests/e2e/batches/batch_client.py:40
↓ 6 callersFunctionbatch_upsert_tools
Write a batch of discovered tools to LiteLLM_ToolTable in a single statement. Raises on failure so the caller decides whether the rows can be retr
litellm/proxy/db/tool_registry_writer.py:119
↓ 6 callersFunctionbedrock_guardrail_cost
(usage_units: Mapping[str, int], aws_region_name: str | None)
litellm/litellm_core_utils/llm_cost_calc/guardrail_cost.py:104
↓ 6 callersFunctionbedrock_route_for_request
The route for one request, decided from the caller's raw params before any provider mapping. Param mapping and dispatch both call this with the s
litellm/llms/bedrock/common_utils.py:1444
↓ 6 callersFunctionbegin
( py: Python<'py>, locals: &Bound<'py, PyDict>, asynchronous: bool, )
litellm-rust/crates/callbacks-legacy-python/src/adapter.rs:733
↓ 6 callersMethodbind_agent_lookup
(self, agent_lookup: AgentLookup)
litellm/proxy/auth/handle_jwt.py:284
↓ 6 callersFunctionbinding
()
tests/test_litellm_rust/test_dispatch.py:19
↓ 6 callersFunctionbinding
(native: NativeFn | None)
tests/unit/rust_bridge/test_runtime.py:35
↓ 6 callersFunctionblock
(name: str, metadata: Mapping[str, JsonValue] | None = METADATA)
tests/integration/caching/test_embedding_file_block_cache.py:35
↓ 6 callersFunctionblock_key
Block an Virtual key from making any requests. Parameters: - key: str - The key to block. Can be either the unhashed key (sk-...) or the
litellm/proxy/management_endpoints/key_management_endpoints.py:8031
↓ 6 callersFunctionblock_texts
(frames: Sequence[Frame])
tests/integration/observability/test_bedrock_converse_stream_post_call_guardrail.py:117
↓ 6 callersFunctionbody
(response: Response)
litellm-rust/crates/gateway-ui/tests/routes.rs:35
↓ 6 callersFunctionbreaker
(failure_threshold: u32)
litellm-rust/crates/redis-client/tests/client.rs:743
↓ 6 callersMethodbreaker
The breaker of the server `endpoint` names, built from `config` by its first caller. Without an endpoint the breaker is this caller's own, which `brea
litellm-rust/crates/redis-connect/src/hosts.rs:27
↓ 6 callersFunctionbridge_client_tool_search
( input: str | ResponseInputParam, tools: Sequence[ALL_RESPONSES_API_TOOL_PARAMS] | None, loaded_e
litellm/responses/litellm_completion_transformation/client_tool_search.py:243
↓ 6 callersFunctionbroadcast_evictions
Tell every other worker to drop these keys, in one pipelined Redis round trip. For bulk callers that already dropped the keys here and from
litellm/proxy/common_utils/auth_cache_invalidation_pubsub.py:164
↓ 6 callersFunctionbudget_cascade_unit_of_work
( new_batch: Callable[[], PrismaBatch], )
litellm/repositories/unit_of_work.py:202
↓ 6 callersFunctionbuild
(make_core: MakeCore, **arguments: object)
tests/unit/test_router/core/test_gaps_l2800_streaming.py:38
↓ 6 callersFunctionbuildComplexityRouterConfig
({ tiers, enableNonReasoningTier, customTierSet, defaultModel, planModeMinTier, tierLabels, clas
ui/litellm-dashboard/src/components/add_model/build_complexity_router_config.ts:653
↓ 6 callersFunctionbuildModelAvailability
( modelGroups: Iterable<string>, deployments: readonly DeploymentModelRef[], groupProviders: ReadonlyMap
ui/litellm-dashboard/src/lib/autorouter_presets.ts:117
↓ 6 callersFunctionbuild_auto_router_catalog
(rows: Sequence[object])
litellm/proxy/management_helpers/auto_router_availability.py:54
↓ 6 callersMethodbuild_block_sse_chunks
Build Responses API SSE events that deliver the guardrail block message and terminate the stream cleanly, mirroring the non-streaming
litellm/llms/openai/responses/guardrail_translation/handler.py:1974
↓ 6 callersFunctionbuild_competitor_guardrail_definitions
Build enriched guardrailDefinitions with competitor names and variations populated.
litellm/proxy/management_endpoints/policy_endpoints/endpoints.py:1023
↓ 6 callersFunctionbuild_jwks
(public_key: "ec.EllipticCurvePublicKey")
litellm/llms/base_llm/auth/jwt_signing.py:85
↓ 6 callersMethodbuild_litellm_proxy_success_headers_from_llm_response
Build LiteLLM proxy response headers for routes that call the LLM directly (e.g. Google native :generateContent) instead of base_proc
litellm/proxy/common_request_processing.py:2480
↓ 6 callersFunctionbuild_proxy_bug_report
( exc: BaseException, *, call_type: str | None = None, custom_llm_provider: object = None,
litellm/proxy/bug_report_config.py:214
↓ 6 callersFunctionbuild_request
(model: &str, image_url: String, params: &CohereOptions)
litellm-rust/crates/llms/src/cohere/ocr/transformation.rs:268
↓ 6 callersFunctionbuild_router_embedding_metadata
Forward the caller's full metadata, flagged as a semantic-cache embedding.
litellm/caching/_embedding_router.py:41
↓ 6 callersMethodbuild_routes
Build semantic-router Route objects from templates + custom config.
litellm/proxy/guardrails/guardrail_hooks/semantic_guard/route_loader.py:69
↓ 6 callersFunctionbuild_spend_event
Validate the success callback's kwargs and serialize the event once, as a single JSON line.
litellm/proxy/spend_tracking/spend_event.py:347
↓ 6 callersFunctionbuild_tool_search_call
( tool_id: str, arguments_json: str, status: ToolSearchCallStatus, execution: ToolSearchExecut
litellm/responses/litellm_completion_transformation/tool_search.py:387
↓ 6 callersFunctionbulk_team_member_add
Bulk add multiple members to a team at once. This endpoint reuses the same logic as /team/member_add but provides a bulk-friendly respon
litellm/proxy/management_endpoints/team_endpoints.py:4608
↓ 6 callersFunctionc
()
litellm/proxy/_experimental/out/_next/static/chunks/1a2lzi3ursud4.js:1
↓ 6 callersFunctionc
(e)
litellm/proxy/_experimental/out/_next/static/chunks/0_6jqgnorqxl2.js:1
↓ 6 callersFunctionc9
(e)
litellm/proxy/_experimental/out/_next/static/chunks/0_6jqgnorqxl2.js:1
↓ 6 callersFunctioncW
()
litellm/proxy/_experimental/out/_next/static/chunks/085r5ieqhajig.js:10
↓ 6 callersFunctioncache_control_as_prompt_cache_breakpoint
Anthropic's ``cache_control`` mark rewritten as OpenAI's ``prompt_cache_breakpoint`` on the same block.
litellm/litellm_core_utils/prompt_templates/prompt_cache_breakpoint_utils.py:93
↓ 6 callersFunctioncache_key
(response: object)
tests/test_litellm_rust/cache/support.py:33
↓ 6 callersFunctioncache_step
( op: CacheOp, key: Option<&str>, keys: Option<&[String]>, value: Option<Value>, ttl: Opti
litellm-rust/crates/python-bridge/src/router/host.rs:556
↓ 6 callersFunctioncached_get_model_info_helper
_get_model_info_helper wrapped with lru_cache Speed Optimization to hit high RPS
litellm/utils.py:7177
↓ 6 callersFunctioncalculate_image_response_cost_from_usage
Calculate image generation cost from usage metadata when available. Returns: Optional[float]: total cost from token usage, or None w
litellm/litellm_core_utils/llm_cost_calc/utils.py:2101
↓ 6 callersMethodcalculate_mcp_tool_call_cost
Calculate the cost of an MCP tool call. Default is 0.0, unless user specifies a custom cost per request for MCP tools.
litellm/proxy/mcp_server/cost_calculator.py:20
↓ 6 callersMethodcalculate_spend
(self, model: str, content: str)
tests/e2e/quota_management/spend_tracking/spend_e2e_client.py:324
↓ 6 callersFunctioncall
(gateway: Gateway, endpoint: Endpoint, deployments: Deployments, text: str)
tests/integration/providers/test_vertex_credentials_source_wire.py:113
↓ 6 callersFunctioncall
(marker: str)
tests/integration/providers/test_api_base_url_join_chaos.py:68
↓ 6 callersFunctioncall
()
tests/integration/providers/test_anthropic_refusal_stop_details_wire.py:296
↓ 6 callersFunctioncall
()
tests/unit/llms/bigmodel/test_bigmodel_provider.py:612
↓ 6 callersFunctioncall
(index: int, city: str)
tests/unit/completion_extras/litellm_responses_transformation/test_completion_extras_litellm_responses_transformation_transformation.py:4184
↓ 6 callersFunctioncall
(provider: &str, payload: Value)
litellm-rust/crates/inference-messages/tests/live/support.rs:87
↓ 6 callersFunctioncall
( cache: &Arc<dyn ResponseCacheService>, options: Option<CacheOptions>, calls: &AtomicUsize, r
litellm-rust/crates/inference/tests/caching.rs:92
↓ 6 callersFunctioncall_decisions
( is_async: bool, systemone: bool, credentials: Mapping[str, object], api_base: str = _BASE, )
tests/unit/llms/azure/decisions/test_transformation.py:51
↓ 6 callersFunctioncall_edit
( harness: Harness, *, body: Dict[str, Any], headers=None, query=None )
tests/unit/proxy/video_endpoints/test_endpoints.py:515
↓ 6 callersMethodcall_later
( self, delay: float, callback: Callable[..., object], *args: object, context: Context | None
tests/unit/llms/custom_httpx/test_http_handler.py:2461
↓ 6 callersFunctioncall_surface
Drive one request through the given surface; return (client-visible response id, x-litellm-call-id).
tests/integration/observability/_s3_v2_support.py:257
↓ 6 callersFunctioncaller
( scenario: Scenario, *, team_id: str | None = None, team_metadata: Mapping[str, JsonValue] |
tests/integration/security/test_callback_credentials.py:113
↓ 6 callersMethodcalls
(self)
tests/integration/compatibility/test_compressed_request_bodies.py:87
↓ 6 callersMethodcalls
(self)
tests/unit/proxy/db/conftest.py:120
↓ 6 callersFunctioncanModifyModel
( actor: ModelActor, teams: Team[] | null, { teamId, isDbModel }: ModelRowOrigin, )
ui/litellm-dashboard/src/utils/modelPermissions.ts:81
↓ 6 callersMethodcan_user_make_team_model_call
( team_id: str, user_api_key_dict: UserAPIKeyAuth, team_obj: LiteLLM_TeamTable | None
litellm/proxy/management_endpoints/model_management_endpoints.py:2660
↓ 6 callersMethodcanonical_model
(self, model: str)
litellm/llms/laya/decisions/transformation.py:10
↓ 6 callersFunctioncanonical_provider_file_id
The one spelling of a provider file id that ownership rows store and look up, so ``files/abc``, ``files%2Fabc``, ``abc/`` and the Gemini file URL
litellm/proxy/openai_files_endpoints/common_utils.py:355
↓ 6 callersFunctioncanonicalize_url_identity
Normalize a URL to a comparable identity: lowercase scheme and host, drop the scheme's default port, and strip userinfo, params, query, fragment a
litellm/proxy/mcp_server/oauth_utils.py:734
↓ 6 callersFunctioncapability_limit_violation
Why holding ``held`` routers claiming ``capability`` exceeds ``limit``, or None when it fits. ``limit`` None means unlimited. The message is shar
litellm/router_strategy/auto_router/auto_router_model_naming.py:362
↓ 6 callersFunctioncapture_serialized_event
(env: Mapping[str, str], raiser: Callable[[], None] = raise_with_identity_locals)
tests/unit/litellm_core_utils/test_sentry_scrubbing.py:75
↓ 6 callersFunctioncassette_cache_capacity_snapshot
Probe Redis ``INFO memory`` and return used/max bytes and percent. Returns ``None`` if Redis is unreachable, the server didn't report ``maxme
tests/_vcr_redis_persister.py:62
↓ 6 callersFunctionce
(e)
litellm/proxy/_experimental/out/_next/static/chunks/0_6jqgnorqxl2.js:1
↓ 6 callersFunctionce
(t)
litellm/proxy/_experimental/out/_next/static/chunks/0cz1d0mv5g_q7.js:1
↓ 6 callersFunctioncharged_rows
(model_id: str)
tests/integration/spend/_background_responses.py:267
↓ 6 callersFunctionchat_completions_route
()
litellm-rust/crates/inference-chat/tests/support/mod.rs:18
↓ 6 callersMethodcheck
(&mut self, row: &Row, what: &str, expected: &str, actual: &str)
litellm-rust/crates/python-compat/tests/fixtures.rs:151
↓ 6 callersFunctioncheck_org_admin_can_generate_keys
Raise HTTP 403 if the caller is an org admin and key generation is disabled for org admins via UI settings. Only blocks the ORG_ADMIN ro
litellm/proxy/common_utils/rbac_utils.py:71
↓ 6 callersFunctioncheck_unsafe_filename
Reject a filename before it can influence any storage path or backend call. Only flags a genuine traversal component ("..") or a null byte, so an
litellm/proxy/openai_files_endpoints/general_upload_validation.py:125
↓ 6 callersFunctionchunk_lines
Group serialized lines into objects of at most ``max_bytes`` uncompressed. A line above the bound on its own still becomes its own object. A
litellm/integrations/pointfive/payload.py:33
↓ 6 callersMethodchunk_parser
Parse individual chunks from streaming response
litellm/llms/ovhcloud/chat/transformation.py:74
↓ 6 callersFunctionchunks_response
(chunks: Value)
litellm-rust/crates/inference-ocr/tests/ocr/reducto.rs:13
↓ 6 callersFunctionclassify_upstream_dcr_rejection
Classify a dynamic-client-registration rejection. RFC 7591 §3.2.2 errors carry ``error`` / ``error_description`` and go through the same blame ass
litellm/proxy/mcp_server/faults/classify.py:125
↓ 6 callersFunctionclaude_routes
(binary: &[u8], checksum: &str)
litellm-rust/crates/testkit/tests/install.rs:33
↓ 6 callersFunctioncleanup
()
ui/litellm-dashboard/src/components/liteadmin/answerExport.ts:226
↓ 6 callersMethodclear
(self)
litellm/proxy/config_resolvers/settings_store.py:136
↓ 6 callersMethodclear_team_member_budget_fields
Clear explicitly-nulled fields on the team member budget row.
litellm/proxy/management_endpoints/team_endpoints.py:875
↓ 6 callersFunctionclock
()
litellm-rust/crates/redis-store/src/dual_cache/mod.rs:134
↓ 6 callersMethodclose
Close the connection and cleanup.
cookbook/nova_sonic_realtime.py:220
↓ 6 callersMethodclose
Flush the buffer for up to ``drain_timeout`` seconds, then apply the unavailable policy to the rest.
litellm/proxy/spend_tracking/spend_event_producer.py:212
↓ 6 callersMethodclose
(self)
litellm/proxy/roi_calculator/github.py:377
← previousnext →12,701–12,800 of 202,479, ranked by callers