Llama Stack Files

What's Changed

• chore: update convert_tooldef_to_openai_tool to match its usage by @mattf in https://github.com/llamastack/llama-stack/pull/4837
• feat!: improve consistency of post-training API endpoints by @eoinfennessy in https://github.com/llamastack/llama-stack/pull/4606
• fix: Arbitrary file write via a non-default configuration by @VaishnaviHire in https://github.com/llamastack/llama-stack/pull/4844
• chore: reduce uses of models.llama.datatypes by @mattf in https://github.com/llamastack/llama-stack/pull/4847
• docs: add technical release steps and improvements to RELEASE_PROCESS.md by @cdoern in https://github.com/llamastack/llama-stack/pull/4792
• chore: bump fallback version to 0.5.1 by @cdoern in https://github.com/llamastack/llama-stack/pull/4846
• fix: Exclude null 'strict' field in function tools to prevent OpenAI … by @gyliu513 in https://github.com/llamastack/llama-stack/pull/4795
• chore(test): add test to verify responses params make it to backend service by @mattf in https://github.com/llamastack/llama-stack/pull/4850
• chore: revert "fix: disable together banner (#4517)" by @mattf in https://github.com/llamastack/llama-stack/pull/4856
• fix: update together to work with latest api.together.xyz service (circa feb 2026) by @mattf in https://github.com/llamastack/llama-stack/pull/4857
• chore(github-deps): bump astral-sh/setup-uv from 7.2.0 to 7.3.0 by @dependabot[bot] in https://github.com/llamastack/llama-stack/pull/4867
• chore(github-deps): bump github/codeql-action from 4.32.0 to 4.32.2 by @dependabot[bot] in https://github.com/llamastack/llama-stack/pull/4861
• chore(github-deps): bump actions/cache from 5.0.2 to 5.0.3 by @dependabot[bot] in https://github.com/llamastack/llama-stack/pull/4859
• chore(github-deps): bump llamastack/llama-stack from [76bcb6] to [c518b3] by @dependabot[bot] in https://github.com/llamastack/llama-stack/pull/4858
• fix(ci): ensure oasdiff is available for openai-coverage hook by @EleanorWho in https://github.com/llamastack/llama-stack/pull/4835
• fix: Deprecate items when create conversation by @gyliu513 in https://github.com/llamastack/llama-stack/pull/4765
• chore: refactor chunking to use configurable tiktoken encoding and document tokenizer limits by @mattf in https://github.com/llamastack/llama-stack/pull/4870
• chore: prune unused parts of models packages (checkpoint, tokenizer, prompt templates, datatypes) by @mattf in https://github.com/llamastack/llama-stack/pull/4871
• chore: prune unused utils from utils.memory.vector_store by @mattf in https://github.com/llamastack/llama-stack/pull/4873
• fix: Escape special characters in auto-generated provider documentati… by @gyliu513 in https://github.com/llamastack/llama-stack/pull/4822
• chore(docs): Use starter for opentelemetry integration test by @gyliu513 in https://github.com/llamastack/llama-stack/pull/4875
• fix: kvstore should call shutdown but not close by @gyliu513 in https://github.com/llamastack/llama-stack/pull/4872
• fix: uvicorn log ambiguity by @cdoern in https://github.com/llamastack/llama-stack/pull/4522
• chore(github-deps): bump actions/checkout from 4.2.2 to 6.0.2 by @dependabot[bot] in https://github.com/llamastack/llama-stack/pull/4865
• chore: cleanup mypy excludes by @mattf in https://github.com/llamastack/llama-stack/pull/4876
• feat: add integration test for max_output_tokens by @gyliu513 in https://github.com/llamastack/llama-stack/pull/4825
• chore(test): add test to verify responses params make it to backend s… by @gyliu513 in https://github.com/llamastack/llama-stack/pull/4852
• ci: add Docker image publishing to release workflow by @cdoern in https://github.com/llamastack/llama-stack/pull/4882
• feat: add ProcessFileRequest model to file_processors API by @alinaryan in https://github.com/llamastack/llama-stack/pull/4885
• docs: update responses api known limitations doc by @jaideepr97 in https://github.com/llamastack/llama-stack/pull/4845
• fix(vector_io): align Protocol signatures with request models by @skamenan7 in https://github.com/llamastack/llama-stack/pull/4747
• fix: add _ExceptionTranslatingRoute to prevent keep-alive breakage on Linux by @iamemilio in https://github.com/llamastack/llama-stack/pull/4886
• docs: add release notes for version 0.5 by @rhuss in https://github.com/llamastack/llama-stack/pull/4855
• fix(ci): disable uv cache cleanup when UV_NO_CACHE is set by @cdoern in https://github.com/llamastack/llama-stack/pull/4889
• feat: Add truncation parameter support by @gyliu513 in https://github.com/llamastack/llama-stack/pull/4813
• chore(ci): bump pinned action commit hashes in integration-tests.yml by @cdoern in https://github.com/llamastack/llama-stack/pull/4895
• docs: Add README for running observability test by @gyliu513 in https://github.com/llamastack/llama-stack/pull/4884
• fix: update rerank routing to match params by @mattf in https://github.com/llamastack/llama-stack/pull/4900
• feat: Add prompt_cache_key parameter support by @gyliu513 in https://github.com/llamastack/llama-stack/pull/4775
• chore: add rerank support to recorder by @mattf in https://github.com/llamastack/llama-stack/pull/4903
• feat: add rerank support to vllm inference provider by @mattf in https://github.com/llamastack/llama-stack/pull/4902
• fix(inference): use flat response message model for chat/completions by @cdoern in https://github.com/llamastack/llama-stack/pull/4891
• feat: add llama cpp server remote inference provider by @Bobbins228 in https://github.com/llamastack/llama-stack/pull/4382
• fix: Remove pillow as direct dependency by @VaishnaviHire in https://github.com/llamastack/llama-stack/pull/4901
• fix: pre-commit run -a by @mattf in https://github.com/llamastack/llama-stack/pull/4907
• fix(ci): Removed kotlin from preview builds by @gyliu513 in https://github.com/llamastack/llama-stack/pull/4910
• feat: Add service_tier parameter support by @gyliu513 in https://github.com/llamastack/llama-stack/pull/4816
• chore(github-deps): bump github/codeql-action from 4.32.2 to 4.32.3 by @dependabot[bot] in https://github.com/llamastack/llama-stack/pull/4918
• chore(github-deps): bump docker/login-action from 3.4.0 to 3.7.0 by @dependabot[bot] in https://github.com/llamastack/llama-stack/pull/4916
• chore(github-deps): bump llamastack/llama-stack from [c7cdb4] to [4c1b03] by @dependabot[bot] in https://github.com/llamastack/llama-stack/pull/4915
• chore(github-deps): bump stainless-api/upload-openapi-spec-action from 1.10.0 to 1.11.6 by @dependabot[bot] in https://github.com/llamastack/llama-stack/pull/4913
• chore(github-deps): bump docker/build-push-action from 6.15.0 to 6.19.2 by @dependabot[bot] in https://github.com/llamastack/llama-stack/pull/4912
• fix(vertexai): raise descriptive error on auth failure instead of silent empty string by @major in https://github.com/llamastack/llama-stack/pull/4909
• fix: resolve StorageConfig default env vars at construction time by @major in https://github.com/llamastack/llama-stack/pull/4897
• feat: Add incomplete_details response property by @gyliu513 in https://github.com/llamastack/llama-stack/pull/4812
• feat(client-sdks): add OpenAPI Generator tooling by @aegeiger in https://github.com/llamastack/llama-stack/pull/4874
• fix(vector_io): eliminate duplicate call for vector store registration by @r3v5 in https://github.com/llamastack/llama-stack/pull/4925
• test(vertexai): add unit tests for VertexAI inference adapter by @major in https://github.com/llamastack/llama-stack/pull/4927
• feat: introduce new how-to blog by @cdoern in https://github.com/llamastack/llama-stack/pull/4794
• chore: remove reference to non-existent WeaviateRequestProviderData by @mattf in https://github.com/llamastack/llama-stack/pull/4937
• feat: standardized error types with HTTP status codes by @iamemilio in https://github.com/llamastack/llama-stack/pull/4878
• feat: add opentelemetry-distro to core dependencies by @Artemon-line in https://github.com/llamastack/llama-stack/pull/4935
• feat(ci): Add nightly job for doc build by @gyliu513 in https://github.com/llamastack/llama-stack/pull/4911
• fix: Ensure user isolation for stored conversations and responses by @jaideepr97 in https://github.com/llamastack/llama-stack/pull/4834
• fix: align chat completion usage schema with OpenAI spec by @cdoern in https://github.com/llamastack/llama-stack/pull/4930
• fix: allow conversation item type to be omitted by @mattf in https://github.com/llamastack/llama-stack/pull/4948
• feat: Enable inline PyPDF file_processors provider by @alinaryan in https://github.com/llamastack/llama-stack/pull/4743
• feat: add support for /responses background parameter by @cdoern in https://github.com/llamastack/llama-stack/pull/4824
• feat(vector_io): Implement Contextual Retrieval for improved RAG search quality by @r-bit-rry in https://github.com/llamastack/llama-stack/pull/4750
• chore: use SecretStr for x-llamastack-provider-data keys by @mattf in https://github.com/llamastack/llama-stack/pull/4939
• chore: remove unused vector store utils by @mattf in https://github.com/llamastack/llama-stack/pull/4961
• feat: auto-identify embedding models for vllm by @mattf in https://github.com/llamastack/llama-stack/pull/4975
• chore(github-deps): bump llamastack/llama-stack from [4c1b03] to [7d9786] by @dependabot[bot] in https://github.com/llamastack/llama-stack/pull/4971
• chore(github-deps): bump actions/checkout from 6.0.1 to 6.0.2 by @dependabot[bot] in https://github.com/llamastack/llama-stack/pull/4969
• chore(github-deps): bump actions/cache from 4.2.0 to 5.0.3 by @dependabot[bot] in https://github.com/llamastack/llama-stack/pull/4963
• chore(github-deps): bump github/codeql-action from 4.32.3 to 4.32.4 by @dependabot[bot] in https://github.com/llamastack/llama-stack/pull/4964
• chore(github-deps): bump actions/stale from 10.1.1 to 10.2.0 by @dependabot[bot] in https://github.com/llamastack/llama-stack/pull/4966
• fix: fix connector_id resolution in agent provider by @jaideepr97 in https://github.com/llamastack/llama-stack/pull/4853
• build: bump fallback_version to 0.5.2.dev0 post 0.5.1 release by @cdoern in https://github.com/llamastack/llama-stack/pull/4959
• fix: pass request objects to Files API in Responses content conversion by @mattf in https://github.com/llamastack/llama-stack/pull/4977
• fix: test_prepend_prompt_with_mixed_variables mock by @mattf in https://github.com/llamastack/llama-stack/pull/4979
• feat: enforce max upload size for Files and File Processors APIs by @alinaryan in https://github.com/llamastack/llama-stack/pull/4956
• feat: add OpenResponses conformance CI job with replay recordings by @cdoern in https://github.com/llamastack/llama-stack/pull/4981
• feat(client-sdks): add hierarchical SDK build pipeline by @aegeiger in https://github.com/llamastack/llama-stack/pull/4932
• feat: add top_p parameter support to responses API by @EleanorWho in https://github.com/llamastack/llama-stack/pull/4820
• fix(docs): Updated llamastack pod metadata by @gyliu513 in https://github.com/llamastack/llama-stack/pull/4983
• chore: move parse_data_url to common package by @mattf in https://github.com/llamastack/llama-stack/pull/4982
• feat: record and replay provider exceptions in inferencing integration tests by @iamemilio in https://github.com/llamastack/llama-stack/pull/4880
• feat: Use Structured Errors in Responses and Conversations API by @iamemilio in https://github.com/llamastack/llama-stack/pull/4879
• fix: strip inline:: prefix from model in vector io tests by @mattf in https://github.com/llamastack/llama-stack/pull/4993
• refactor: consolidate dynamic provider config parsing by @mattf in https://github.com/llamastack/llama-stack/pull/4985
• feat: auto-merge PRs on stable release branches via Mergify + CI gate by @leseb in https://github.com/llamastack/llama-stack/pull/4992
• refactor: use OpenAIErrorResponse model for consistent error responses by @iamemilio in https://github.com/llamastack/llama-stack/pull/4883
• fix: populate required OpenResponses fields with non-null defaults by @cdoern in https://github.com/llamastack/llama-stack/pull/4994
• feat: auto-merge dependabot github-deps PRs via Mergify by @leseb in https://github.com/llamastack/llama-stack/pull/4995
• feat: Add top_logprobs parameter support by @gyliu513 in https://github.com/llamastack/llama-stack/pull/4814
• feat: add support for 'frequency_penalty' param to Responses API by @nathan-weinberg in https://github.com/llamastack/llama-stack/pull/4823
• feat: add support for 'presence_penalty' param to Responses API by @nathan-weinberg in https://github.com/llamastack/llama-stack/pull/4830
• fix: correct PYPDF adapter method signature to match FileProcessors protocol by @alinaryan in https://github.com/llamastack/llama-stack/pull/4998
• fix(responses): achieve full OpenResponses conformance by @cdoern in https://github.com/llamastack/llama-stack/pull/4999
• fix(docs): Updated health check endpoint by @gyliu513 in https://github.com/llamastack/llama-stack/pull/5000
• test: Add responses structured output integration tests by @msager27 in https://github.com/llamastack/llama-stack/pull/4940
• feat: structured error handling in Responses API streaming by @iamemilio in https://github.com/llamastack/llama-stack/pull/4942
• feat(client-sdks): add LlamaStackClient, httpx, and streaming by @aegeiger in https://github.com/llamastack/llama-stack/pull/5001
• feat: accept list content blocks in Responses API function_call_output by @mattf in https://github.com/llamastack/llama-stack/pull/4978
• refactor(PGVector): wrap gin index creation into a separate function by @r3v5 in https://github.com/llamastack/llama-stack/pull/4980
• chore: consolidate backend-forwarded param tests into unified parametrized test by @mattf in https://github.com/llamastack/llama-stack/pull/5003
• test: add integration tests for Responses and Conversations API errors by @iamemilio in https://github.com/llamastack/llama-stack/pull/4881
• feat: allow stream usage from ollama when telemetry enabled by @mattf in https://github.com/llamastack/llama-stack/pull/5011
• feat: allow stream usage from vllm when telemetry enabled by @mattf in https://github.com/llamastack/llama-stack/pull/5010
• fix(registry): loosen register() idempotent checks for server restarts by @max-svistunov in https://github.com/llamastack/llama-stack/pull/4976
• feat: add integration test for prompt_cache_key with openai client by @gyliu513 in https://github.com/llamastack/llama-stack/pull/5016
• chore(github-deps): bump actions/github-script from 7.0.1 to 8.0.0 by @dependabot[bot] in https://github.com/llamastack/llama-stack/pull/5025
• chore(github-deps): bump astral-sh/setup-uv from 7.3.0 to 7.3.1 by @dependabot[bot] in https://github.com/llamastack/llama-stack/pull/5027
• chore(github-deps): bump actions/setup-java from 4.5.0 to 5.2.0 by @dependabot[bot] in https://github.com/llamastack/llama-stack/pull/5019
• ci: add merge_group trigger to all PR-gating workflows by @cdoern in https://github.com/llamastack/llama-stack/pull/5017
• feat(ci): automate post-release and pre-release version management by @cdoern in https://github.com/llamastack/llama-stack/pull/4938
• test: Add prompt template test cases to the responses integraton test… by @msager27 in https://github.com/llamastack/llama-stack/pull/4950
• fix(stainless): handle [DONE] SSE terminator in streaming responses by @dtmeadows in https://github.com/llamastack/llama-stack/pull/5012
• fix(security): pin google-cloud-aiplatform to >=1.131.0 by @derekhiggins in https://github.com/llamastack/llama-stack/pull/5037
• feat(inference): bidirectional reasoning token passthrough for chat completions by @cdoern in https://github.com/llamastack/llama-stack/pull/5038
• chore: remove unreachable tool_choice check in vllm adapter by @mattf in https://github.com/llamastack/llama-stack/pull/5009
• feat(api): support extra_body pass-through in responses API by @codefromthecrypt in https://github.com/llamastack/llama-stack/pull/4893
• docs: additional references to Docker Hub by @nathan-weinberg in https://github.com/llamastack/llama-stack/pull/5044
• fix: add missing shutdown method to PyPDF file processor adapter by @alinaryan in https://github.com/llamastack/llama-stack/pull/5047
• fix(llama-guard): less strict parsing of safety categories by @asimurka in https://github.com/llamastack/llama-stack/pull/5045
• fix: OCI26ai sql query patches by @rhdedgar in https://github.com/llamastack/llama-stack/pull/5046
• fix(conversations): validate conv_ prefix consistently on all endpoints by @iamemilio in https://github.com/llamastack/llama-stack/pull/5058
• fix(conversations): add ExceptionTranslatingRoute to conversations router by @iamemilio in https://github.com/llamastack/llama-stack/pull/5057
• feat: allow model registration without provider API keys by @NickGagan in https://github.com/llamastack/llama-stack/pull/5014
• chore: Rename test_openai_response.py to test_openai_responses.py by @gyliu513 in https://github.com/llamastack/llama-stack/pull/5061
• fix: (pypdf) Possible infinite loop when loading circular /Prev entries in cross-reference streams by @eoinfennessy in https://github.com/llamastack/llama-stack/pull/5063
• chore: bump fallback_version to 0.5.3.dev0 after 0.5.2 release by @cdoern in https://github.com/llamastack/llama-stack/pull/5065
• feat: passthrough safety provider for forwarding to downstream /v1/moderations by @skamenan7 in https://github.com/llamastack/llama-stack/pull/5004
• feat: add conditional authentication provider configuration by @derekhiggins in https://github.com/llamastack/llama-stack/pull/5002
• fix: NLTK Zip Slip Vulnerability by @eoinfennessy in https://github.com/llamastack/llama-stack/pull/5062
• fix: use semantic JSON comparison for MCP approval argument matching by @iamemilio in https://github.com/llamastack/llama-stack/pull/5080
• feat(vertexai): rewrite provider on google-genai with dynamic model listing by @major in https://github.com/llamastack/llama-stack/pull/4951
• ci: temporarily disable CodeQL workflow on pull requests by @leseb in https://github.com/llamastack/llama-stack/pull/5079
• chore: fix post-release workflow and remove broken docker image by @cdoern in https://github.com/llamastack/llama-stack/pull/5064
• fix: Revert "temporarily disable CodeQL workflow on pull requests" by @cdoern in https://github.com/llamastack/llama-stack/pull/5085
• fix: use canonical config loading in backward compat test by @leseb in https://github.com/llamastack/llama-stack/pull/5081
• fix!: add content capture via otel by @gyliu513 in https://github.com/llamastack/llama-stack/pull/5060
• feat!: add integration test for safety_identifier with openai client by @gyliu513 in https://github.com/llamastack/llama-stack/pull/5018
• fix: poll test PyPI before building Docker images to avoid race condition by @cdoern in https://github.com/llamastack/llama-stack/pull/5090
• feat: add regex pattern support to access_policy and route_policy by @derekhiggins in https://github.com/llamastack/llama-stack/pull/4991
• feat: add integration test for truncation with openai client by @gyliu513 in https://github.com/llamastack/llama-stack/pull/5084
• feat: integration test for top_p with openai client by @gyliu513 in https://github.com/llamastack/llama-stack/pull/5083
• chore(github-deps): bump oven-sh/setup-bun from 2.1.2 to 2.1.3 by @dependabot[bot] in https://github.com/llamastack/llama-stack/pull/5068
• chore(github-deps): bump github/codeql-action from 4.32.4 to 4.32.6 by @dependabot[bot] in https://github.com/llamastack/llama-stack/pull/5070
• chore(github-deps): bump llamastack/llama-stack from [7d9786] to [6c700d] by @dependabot[bot] in https://github.com/llamastack/llama-stack/pull/5075
• chore(github-deps): bump actions/setup-node from 6.2.0 to 6.3.0 by @dependabot[bot] in https://github.com/llamastack/llama-stack/pull/5067
• chore(github-deps): bump actions/download-artifact from 7.0.0 to 8.0.0 by @dependabot[bot] in https://github.com/llamastack/llama-stack/pull/5020
• fix: process hang on exit with aiosqlite >= 0.22 by @shanemcd in https://github.com/llamastack/llama-stack/pull/4589
• fix: treat hallucinated tool names as client-side function calls by @mattf in https://github.com/llamastack/llama-stack/pull/5043
• test: add streaming web_search test cases to responses integration test suite by @msager27 in https://github.com/llamastack/llama-stack/pull/4960
• feat(PGVector): implement ef_search parameter for HNSW vector index in PGVector by @r3v5 in https://github.com/llamastack/llama-stack/pull/4933
• feat: improve inference performance via cached ssl context by @mattf in https://github.com/llamastack/llama-stack/pull/4486
• ci: update Mergify config with auto-update, auto-approve, and fix merge method by @leseb in https://github.com/llamastack/llama-stack/pull/5091
• feat: Add additional Tool runtime metrics by @gyliu513 in https://github.com/llamastack/llama-stack/pull/4904
• feat: Add integration test for parallel_tool_calls with openai client by @gyliu513 in https://github.com/llamastack/llama-stack/pull/5093
• feat: Enable Filters in OpenAI Search API by @franciscojavierarceo in https://github.com/llamastack/llama-stack/pull/4471
• chore: Move background integration test to test_open_responses.py by @gyliu513 in https://github.com/llamastack/llama-stack/pull/5094
• refactor(vertexai): extract network helper functions into utils module by @major in https://github.com/llamastack/llama-stack/pull/5095
• feat!: new URL for AWS Bedrock and model list support by @are-ces in https://github.com/llamastack/llama-stack/pull/4946

New Contributors

• @major made their first contribution in https://github.com/llamastack/llama-stack/pull/4909
• @aegeiger made their first contribution in https://github.com/llamastack/llama-stack/pull/4874
• @Artemon-line made their first contribution in https://github.com/llamastack/llama-stack/pull/4935
• @max-svistunov made their first contribution in https://github.com/llamastack/llama-stack/pull/4976
• @dtmeadows made their first contribution in https://github.com/llamastack/llama-stack/pull/5012
• @NickGagan made their first contribution in https://github.com/llamastack/llama-stack/pull/5014
• @shanemcd made their first contribution in https://github.com/llamastack/llama-stack/pull/4589

Full Changelog: https://github.com/llamastack/llama-stack/compare/v0.5.2...v0.6.0