* fix(opentelemetry.py): fix issue where headers were not being split correctly
* feat(bedrock/image): Support bedrock titan image generation
Closes https://github.com/BerriAI/litellm/issues/361
* build(model_prices_and_context_window.json): track titan image gen pricing
enables cost tracking per request
* feat(amazon_titan_transformation.py): support titan image generation cost tracking
* docs: document new model
* docs: update docs to indicate cost tracking + refactor rerank into separate doc
* fix: fix mypy linting error
* fix: fix type ignore
* feat(responses_id_security.py): encrypt response.id - prevent user A from retrieving user B's response
additional security for retrievals on shared accounts
Closes LIT-1307
* feat(responses_id_security.py): allow admin to disable responses id security check
* test: add initial unit testing
* feat(responses_id_security.py): add streaming support
* docs: document new param
* docs: document new param
* feat(responses_id_security.py): add team id checks - ensure it works for service accounts
prevent service accounts keys from different teams from accessing each other's responses
more secure
* test: add unit testing
* fix: fix linting error
* fix(presidio.py): handle content as a list of texts
covers openai + anthropic messages api
* fix(presidio.py): safe get messages
* test: add unit testing for presidio guardrails
* fix(unified_guardrail.py): initial commit
* fix(enkryptai.py): implement apply_guardrail to enkrypt guardrail
* fix(unified_guardrail.py): support unified guardrail on input
* feat(unified_guardrail.py): add post call success hook implementation
allows us to just have 1 place to handle llm translation to guardrail api spec
* refactor: refactor initial unified guardrail component
* refactor: more refactoring
* feat(responses/): add guardrails to responses api
allows existing guardrails to work for new llm endpoints
* docs(adding_guardrail_support.md): document new guardrail endpoint support
* test: add unit tests
* feat(image_generation/): add guardrail support for image generation endpoint
* feat(openai/text_completion): support guardrails on `/v1/completions` API
* docs: document guardrails support on new endpoints
* docs: clarify when guardrails run
* feat(openai/speech): add guardrail support for input
* docs(rerank/): add guardrail support on input query
* fix: fix ruff check
* feat(vector_stores/): initial commit adding Vertex AI Search API support for litellm
new vector store provider
* feat(vector_store/): use vector store id for vertex ai search api
* fix: transformation.py
cleanup
* fix: implement abstract function
* fix: fix linting error
* fix: main.py
fix check
* feat: initial commit with working passthrough support for vertex ai search api through litellm
* feat(llm_passthrough_endpoints.py): fix passing correct project on datastore passthrough
* feat(vertex_ai/): support passthrough call for vertex ai search vector store
* docs(vertex_ai_search_datastore.md): document new vertex ai passthrough endpoint
* docs(sidebars.js): document new endpoint
* feat: initial commit adding logging for vertex ai passthrough api
allows vertex ai vector search api to work with cost calculation
* feat(vertex_ai/): search vector store cost tracking
* fix(vertex_passthrough_logging_handler.py): log the cost
* fix: improve logged response
* fix(vertex_passthrough_logging_handler.py): logging
* feat(litellm_logging): main.py
add cost tracking for vertex ai search api via unified api
* refactor: fix ruff checks
* fix(llm_passthrough_endpoints.py): fix linting
* fix(managed_files.py): don't raise error if managed object is not found
* feat(vector_stores): add azure ai search vector store support
Enables direct querying a vector store on azure
* fix(azure/vector_stores): working azure ai search api vector stores
allows azure direct querying on vector stores
* test: update env vars
* docs(docs/): document new azure ai vector store search
* docs(azure_ai_vector_stores.md): add table
* docs: clarify support for 'create' vector stores
* fix(vector_stores/endpoints.py): Fixes https://github.com/BerriAI/litellm/issues/14606
* fix: fix linting errors
* fix(oldteams.tsx): allow org admin to create team on ui
* fix(oldteams.tsx): show org admin a dropdown of allowed orgs for team creation
* docs(access_control.md): cleanup doc
* feat(ibm_guardrails/): initial commit adding support for ibm guardrails on litellm
allows user to use self-hosted ibm guardrails
* feat(ibm_detector.py): working detector
* docs(ibm_guardrails.md): document new ibm guardrails
* fix: fix linting errors
* Addd v2/chat support for cohere
* fix streaming
* Use v2_transformation for logging passthrough:
* Use v2_transformation for logging passthrough:
* Add test for checking if document and citation_options is getting passed
* Update the cohere model
* Add cost tracking for vertex ai passthrough batch jobs
* Add full passthrough support
* refactor code according to the comments
* Add passthrough handler
* remove invalid params
* Updated documentation
* Updated documentation
* Updated documentation
* Correct the import
* Add openai videos generation and retrieval support
* add retrieval endpoint
* Add docs
* Add imports
* remove orjson
* remove double import
* fix openai videos format
* remove mock code
* remove not required comments
* Add tests
* Add tests
* Add other video endpoints
* Fix cost calculation and transformation
* Fixed mypy tests
* remove not used imports
* fix documentation for get batch req (#15742)
* Add grounding info to responses API (#15737)
* Add grounding info to responses API
* fix lint errors
* Use typed objects for annotations
* Use typed objects for annotations
* fix mypy error
* Litellm fix json serialize alreting 2 (#15741)
* fix json serializable error for alerts
* Add test
* fix mypt errors
* fix mypt errors
* Add Qwen3 imported model support for AWS Bedrock (#15783)
* Add qwen imported model support
* fix mypy errors
* fix empty user message error (#15784)
* fix typed dict for list
* Add azure supported videos endpoint
* fix mapped tests
* add azure sora models to model map
* Add OpenAI video generation and content retrieval support (#15745)
* Add openai videos generation and retrieval support
* add retrieval endpoint
* Add docs
* Add imports
* remove orjson
* remove double import
* fix openai videos format
* remove mock code
* remove not required comments
* Add tests
* Add tests
* Add other video endpoints
* Fix cost calculation and transformation
* Fixed mypy tests
* remove not used imports
* fix typed dict for list
* fix mypy errors
* move directory
* make v2 chat default
* Fix mypy tests
* Fix mypy tests
* Fix mypy tests
* Fix mypy tests
* Revert "Add Azure Video Generation Support with Sora Integration"
* refactor videos repo
* add test
* Add azure openai videos support
* Add azure openai videos support
* Add router endpoint support for videos
* fix mypy error
* add azure models
* fix mapped test
* fix mypy error
* Add proxy router test
* Add proxy router test
* remove deprecated model name from tests
* fix import error
* fix import error
* Add gaurdrail integration in videos endpoint
* Add logging support for videos endpoint
* Add final documentation supporting videos integration
* fix model name and document input
* Update literals to avoid mypy errors
* Remove unused imports and print statements
* revert guardrail support for video generation and video remix
* revert guardrail support for video generation and video remix
* Fix failing mapped and llm translation tests
* 1. add v3 classify
2. add new classifix for masking
3. support same id for the conversation for pre and post
working with duplicates
* clean code, remove some debug and run tests
* update liter errors
* improvment for Code Organization, httpx Error Handling Specificity, Logging Improvements and Type
* transfer test test_lasso_guard_config to the new location
* Fix type hints and linting errors in lasso.py
- Add type: ignore for httpx module when None
- Fix return type issues in _handle_classification and _handle_masking
- Ensure masked_messages is not None before passing to _apply_masking_to_model_response
- Convert LassoResponse to dict for _log_masking_applied call
* [Feat] Add SENTRY_ENVIRONMENT configuration for Sentry integration and corresponding tests
* [Refactor] Enhance test_sentry_environment by mocking sentry_sdk and improving environment handling
* [Fix] Update default SENTRY_ENVIRONMENT to 'production' and enhance test for Sentry integration
* [Fix] Update test_sentry_environment to verify correct handling of SENTRY_ENVIRONMENT values
* [Fix] Update test_sentry_environment to assert correct handling of production environment
* feat(guardrails): Add content masking and streaming support to PANW Prisma AIRS
- Add mask_request_content and mask_response_content parameters
- Implement content masking for prompts and responses
- Add streaming support with real-time masking
- Add comprehensive test coverage (28 tests)
- Update documentation with masking examples and security notes
* fix(guardrails): Fix PANW Prisma AIRS env var fallback and text completion support
* perf(router): Optimize prompt management model check with early exit
Add early return for models without '/' to avoid expensive get_model_list()
calls for 99% of standard model requests (gpt-4, claude-3, etc).
- Refactor _is_prompt_management_model() with "/" check before model lookup
- Add unit tests to verify optimization doesn't break detection
* perf(caching): optimize Redis batch cache operations and reduce unnecessary queries
This commit introduces several performance optimizations to the Redis caching layer:
**DualCache Improvements (dual_cache.py):**
1. Increase batch cache size limit from 100 to 1000
- Allows for larger batch operations, reducing Redis round-trips
2. Throttle repeated Redis queries for cache misses
- Update last_redis_batch_access_time for ALL queried keys, including those
with None values
- Prevents excessive Redis queries for frequently-accessed non-existent keys
3. Add early exit optimization
- Short-circuit when redis_result is None or contains only None values
- Avoids unnecessary processing when no cache hits are found
4. Optimize key lookup performance
- Replace O(n) keys.index() calls with O(1) dict lookup via key_to_index mapping
- Reduces algorithmic complexity in batch operations
5. Streamline cache updates
- Combine result updates and in-memory cache updates in single loop
- Only cache non-None values to avoid polluting in-memory cache
**CooldownCache Improvements (cooldown_cache.py):**
1. Enhanced early return logic
- Check if all values in results are None, not just if results is None
- Prevents unnecessary iteration when no valid cooldown data exists
These changes significantly improve Redis caching performance, especially for:
- High-throughput batch operations
- Scenarios with frequent cache misses
- Large-scale deployments with many concurrent requests
* fix: remove unnecessary test
* refactor: move default_max_redis_batch_cache_size to constants
- Add DEFAULT_MAX_REDIS_BATCH_CACHE_SIZE constant (default: 1000)
- Update DualCache to use constant from constants.py
- Document new environment variable in config_settings.md
* fix: only use in memory cache when set
* fix(router): improve prompt management model detection with smart early return
The previous early return optimization in _is_prompt_management_model() was
checking if the model name parameter contained '/' and returning False if it
didn't. This broke detection for model aliases (e.g., 'chatbot_actions') that
don't have '/' in their name but map to prompt management models
(e.g., 'langfuse/openai-gpt-3.5-turbo').
Changed the early return logic to only exit early when:
- Model name contains '/' AND
- The prefix is NOT a known prompt management provider
This maintains the performance optimization for 99% of direct model calls
(avoiding expensive get_model_list lookups) while correctly handling:
- Direct prompt management calls (e.g., 'langfuse/model')
- Model aliases without '/' (e.g., 'chatbot_actions')
- Regular models with/without '/' (e.g., 'gpt-3.5-turbo', 'openai/gpt-4')
Fixes test: test_router_prompt_management_factory
* perf(router): optimize _pre_call_checks with shallow copy (1400x faster)
Replace deepcopy with list() in _pre_call_checks - runs on every request.
Only pops from list, never modifies deployment dicts, so shallow copy is safe.
Performance: 1400x faster on hot path
Impact: 2-5x overall throughput improvement for routing workloads
Tests: Added regression test to ensure no mutation + filtering works
* perf(router): replace deepcopy with shallow copy for default deployment
Replace expensive copy.deepcopy() with shallow copy for default_deployment
in _common_checks_available_deployment() hot path.
Changes:
- Use dict.copy() for top-level deployment dict
- Use dict.copy() for nested litellm_params dict
- Only the 'model' field is modified, so deep recursion is unnecessary
Impact:
- 100x+ faster for default deployment path (every request when used)
- deepcopy recursively traverses entire object tree
- Shallow copy only copies two dict levels (exactly what's needed)
Test coverage:
- Added regression test to verify deployment isolation
- Ensures returned deployments don't mutate original default_deployment
- Validates multiple concurrent requests get independent copies
* perf(router): remove unnecessary dict copy in completion hot paths
Remove unnecessary deployment['litellm_params'].copy() in _completion
and _acompletion functions. The dict is only read and spread into a new
dict, never modified, making the defensive copy wasteful.
Changes:
- Remove .copy() in _completion (sync hot path)
- Remove .copy() in _acompletion (async hot path)
Impact:
- Every completion request (highest traffic endpoints)
- Eliminates unnecessary dict allocation and copy on every call
- Dict spreading already creates new dict, so no mutation possible
Test coverage:
- Added tests verifying deployment params unchanged after calls
- Tests both sync and async completion paths
- Validates optimization doesn't introduce mutations
* perf(router): optimize deployment filtering in pre-call checks
Replace O(n²) list pop pattern with O(n) set-based filtering in
_pre_call_checks() to improve routing performance under high load.
Changes:
- Use set() instead of list for invalid_model_indices tracking
- Replace reversed list.pop() loop with single-pass list comprehension
- Eliminate redundant list→set conversion overhead
Impact:
- Hot path optimization: runs on every request through the router
- ~2-5x faster filtering when many deployments fail validation
- Most beneficial with 50+ deployments per model group or high
invalidation rates (rate limits, context window exceeded)
Technical details:
Old: O(k²) where k = invalid deployments (pop shifts remaining elements)
New: O(n) single pass with O(1) set membership checks
* add: memory profiler
feat(proxy): Add configurable GC thresholds and enhance memory debugging endpoints
- Add PYTHON_GC_THRESHOLD env var to configure garbage collection thresholds
- Add POST /debug/memory/gc/configure endpoint for runtime GC tuning
- Enhance memory debugging endpoints with better structure and explanations
- Add comprehensive router and cache memory tracking
- Include worker PID in all debug responses for multi-worker debugging
* refactor: reduce complexity in get_memory_details endpoint
Extract 6 helper functions from get_memory_details to fix linter
error PLR0915 (too many statements). Improves maintainability
while preserving functionality.
* fix(router): remove incorrect early exit in _is_prompt_management_model
Removes early exit optimization that checked model_name prefix instead
of the actual litellm_params model. This incorrectly returned False for
custom model aliases that map to prompt management providers.
Example: "my-langfuse-prompt/test_id" -> "langfuse_prompt/actual_id"
The method now correctly checks the underlying model's prefix.
Fixes test_is_prompt_management_model_optimization
* fix(proxy): add explicit type annotations to debug_utils dictionaries
Resolved 6 mypy type errors in proxy/common_utils/debug_utils.py by adding
explicit Dict[str, Any] annotations to dictionary variables where mypy was
incorrectly inferring narrow types. This allows the dictionaries to accept
different value types (strings, nested dicts) for error handling and various
return structures.
Fixed:
- Line 246: caches dictionary in get_memory_summary()
- Line 371: cache_stats dictionary in _get_cache_memory_stats()
- Line 439: litellm_router_memory dictionary in _get_router_memory_stats()
* fix(proxy): fix Python 3.8 compatibility in debug_utils type annotations
- Replace tuple[...], list[...] with Tuple[...], List[...] from typing
- Replace Dict | None with Optional[Dict] for Python 3.8 compatibility
- Add missing imports: List, Optional, Tuple to typing imports
Fixes TypeError: 'type' object is not subscriptable in Python 3.8
---------
Co-authored-by: AlexsanderHamir <alexsanderhamirgomesbaptista@gmail.com>
* fix: minor fixes to mcp streaming with bedrock
* fix(bedrock/): working bedrock with mcp tools
handle empty description
* test: add unit test
* test: test fixes
* fix(vector_store_registry.py): load vector store with litellm params from config.yaml
fixes minor issue where litellm params weren't being loaded in from config.yaml
* docs(knowledgebase.md): document azure vector store current limitation
* fix(opentelemetry.py): add hidden params to otel logs
Fixes LIT-1274
* fix: fix test
