Releases: llamastack/llama-stack-client-python
Releases · llamastack/llama-stack-client-python
v0.7.0-alpha.2
0.7.0-alpha.2 (2026-04-01)
Full Changelog: v0.7.0-alpha.1...v0.7.0-alpha.2
Features
- add reasoning as valid conversation item (029da3f)
- add reasoning output types to OpenAI Responses API spec (3bb043e)
Chores
Refactors
- remove deprecated register/unregister model endpoints (6c82145)
v0.7.0-alpha.1
0.7.0-alpha.1 (2026-03-28)
Full Changelog: v0.6.1-alpha.1...v0.7.0-alpha.1
⚠ BREAKING CHANGES
- eliminate GET /chat/completions/{completion_id} conformance issues
- rename agents API to responses API
- eliminate /files/{file_id} GET differences
Features
- Add stream_options parameter support (b4c2f15)
- eliminate /files/{file_id} GET differences (1f28d73)
- eliminate GET /chat/completions/{completion_id} conformance issues (dad9f54)
- internal: implement indices array format for query and form serialization (6694121)
- responses: add cancel endpoint for background responses (d9bc91a)
Bug Fixes
- deps: bump minimum typing-extensions version (50ea4d7)
- inference: improve chat completions OpenAI conformance (147b88b)
- pydantic: do not pass
by_aliasunless set (f6836f9) - remove duplicate dataset_id parameter in append-rows endpoint (d6a79d0)
- sanitize endpoint path params (9b288d5)
Chores
- ci: skip lint on metadata-only changes (b096c2c)
- internal: tweak CI branches (1df7e26)
- internal: update gitignore (0e98cfd)
- internal: version bump (f468096)
- tests: bump steady to v0.19.4 (f5ad8f8)
- tests: bump steady to v0.19.5 (55689e1)
- tests: bump steady to v0.19.6 (87cb87e)
- tests: bump steady to v0.19.7 (10f6ed7)
Refactors
v0.6.1-alpha.1
0.6.1-alpha.1 (2026-03-13)
Full Changelog: v0.5.0-alpha.2...v0.6.1-alpha.1
⚠ BREAKING CHANGES
- improve consistency of post-training API endpoints
Features
- accept list content blocks in Responses API function_call_output (f6f1fc3)
- Add prompt_cache_key parameter support (6b45699)
- add skip_model_availability to openai_mixin for remote models (7ef952b)
- add support for 'frequency_penalty' param to Responses API (56d39cc)
- add support for 'presence_penalty' param to Responses API (4f57d15)
- add support for /responses background parameter (4f8bf45)
- Add top_logprobs parameter support (2196986)
- add top_p parameter support to responses API (23e3b9f)
- Add truncation parameter support (7501365)
- improve consistency of post-training API endpoints (99057fd)
- inference: bidirectional reasoning token passthrough for chat completions (c314639)
- vector_io: Implement Contextual Retrieval for improved RAG search quality (89ec5a7)
Bug Fixes
- align chat completion usage schema with OpenAI spec (3974d5d)
- Enabled models list works (#314) (acd5e64)
- inference: use flat response message model for chat/completions (e58e2e4)
- responses: achieve full OpenResponses conformance — 6/6 tests passing (631ab2c)
- stainless: handle [DONE] SSE terminator in streaming responses (17f0029)
- vector_io: align Protocol signatures with request models (ea58fd8)
Chores
- api: minor updates (17a2705)
- ci: bump uv version (f014d4c)
- ci: skip uploading artifacts on stainless-internal branches (dbddad9)
- docs: add missing descriptions (f1a093b)
- format all
api.mdfiles (0e3e262) - internal: add request options to SSE classes (2ecc682)
- internal: bump dependencies (612291e)
- internal: fix lint error on Python 3.14 (a0f6975)
- internal: make
test_proxy_environment_variablesmore resilient (6bc2bb4) - internal: make
test_proxy_environment_variablesmore resilient to env (44bbae1) - test: do not count install time for mock server timeout (185de33)
- update mock server docs (92cb087)
- update placeholder string (406b9bb)
Refactors
- types: use
extra_itemsfrom PEP 728 (629ca09)
v0.5.0-alpha.2
0.5.0-alpha.2 (2026-02-05)
Full Changelog: v0.5.0-alpha.1...v0.5.0-alpha.2
Features
- Adds support for the
safety_identifierparameter (f20696b)
v0.5.0-alpha.1
0.5.0-alpha.1 (2026-02-04)
Full Changelog: v0.4.0-alpha.15...v0.5.0-alpha.1
Features
- client: add custom JSON encoder for extended type support (ab99ad0)
- client: add support for binary request streaming (d17dede)
Chores
v0.4.0-alpha.15
v0.4.0-alpha.14
0.4.0-alpha.14 (2025-12-22)
Full Changelog: v0.4.0-alpha.13...v0.4.0-alpha.14
Features
- Add metadata field to request and response (a83a98e)
- add support for tool_choice to repsponses api (32c453f)
- api: add readonly connectors API (05f867b)
- api: deprecate
toolgroupandtool_runtimeapis (d5cef5a) - convert Benchmarks API to use FastAPI router (c784ae1)
- convert Datasets API to use FastAPI router (2f53a3f)
- Implement FastAPI router system (2ca4485)
- Implement include parameter specifically for adding logprobs in the output message (05a0330)
- introduce /admin API for stack administration and operations (3279ec9)
- migrate Inspect API to FastAPI router (7774bec)
- migrate Providers API to FastAPI router pattern (f8c84cd)
Bug Fixes
- client: fix issue with duplicate definitions in Go (dd94553)
- ensure streams are always closed (9dc1025)
- types: allow pyright to infer TypedDict types within SequenceNotStr (93be516)
- use async_to_httpx_files in patch method (77df10d)
Chores
- add missing docstrings (18c54f3)
- bump required
uvversion (2a1523b) - deps: mypy 1.18.1 has a regression, pin to 1.17 (a900167)
- internal: add
--fixargument to lint script (89e14ad) - internal: add missing files argument to base client (9cae3df)
- internal: avoid using unstable Python versions in tests (36336dd)
- internal: version bump (597c7c6)
- update lockfile (de67c1e)
- update uv.lock (95cb22a)
Documentation
- add more examples (d24dff6)
v0.4.0-alpha.13
0.4.0-alpha.13 (2025-12-15)
Full Changelog: v0.4.0-alpha.12...v0.4.0-alpha.13
Bug Fixes
- ci: restore push triggers for Stainless preview builds (71c9874)
Chores
v0.4.0-alpha.12
v0.4.0-alpha.11
0.4.0-alpha.11 (2025-11-18)
Full Changelog: v0.4.0-alpha.10...v0.4.0-alpha.11
Features
- Wire through parallel_tool_calls to Responses API (2888bed)