v0.2.19-alpha.1

0.2.19-alpha.1 (2025-08-22)

Full Changelog: v0.2.18-alpha.3...v0.2.19-alpha.1

Features

• api: manual updates (119bdb2)
• api: query_metrics, batches, changes (c935c79)
• api: some updates to query metrics (8f0f7a5)

Bug Fixes

• agent: fix wrong module import in ReAct agent (#262) (c17f3d6), closes #261
• build: kill explicit listing of python3.13 for now (5284b4a)

Chores

• update github action (af6b97e)

Build System

• Bump version to 0.2.18 (53d95ba)

v0.2.18-alpha.3

0.2.18-alpha.3 (2025-08-14)

Full Changelog: v0.2.18-alpha.2...v0.2.18-alpha.3

Features

• llama-stack-client providers inspect PROVIDER_ID (#181) (6d18aae)
• add client-side utility for getting OAuth tokens simply (#230) (91156dc)
• add client.chat.completions.create() and client.completions.create() (#226) (ee0e65e)
• Add llama-stack-client datasets unregister command (#222) (38cd91c)
• add support for chat sessions (#167) (ce3b30f)
• add type hints to event logger util (#140) (26f3c33)
• add updated batch inference types (#220) (ddb93ca)
• add weighted_average aggregation function support (#208) (b62ac6c)
• agent: support multiple tool calls (#192) (43ea2f6)
• agent: support plain function as client_tool (#187) (2ec8044)
• api: update via SDK Studio (48fd19c)
• async agent wrapper (#169) (fc9907c)
• autogen llama-stack-client CLI reference doc (#190) (e7b19a5)
• client.responses.create() and client.responses.retrieve() (#227) (fba5102)
• datasets api updates (#203) (b664564)
• enable_persist: sync updates from stainless branch: yanxi0830/dev (#145) (59a02f0)
• new Agent API (#178) (c2f73b1)
• support client tool output metadata (#180) (8e4fd56)
• Sync updates from stainless branch: ehhuang/dev (#149) (367da69)
• unify max infer iters with server/client tools (#173) (548f2de)
• update react with new agent api (#189) (ac9d1e2)

Bug Fixes

• llama-stack-client provider inspect should use retrieve (#202) (e33b5bf)
• accept extra_headers in agent.create_turn and pass them faithfully (#228) (e72d9e8)
• added uv.lock (546e0df)
• agent: better error handling (#207) (5746f91)
• agent: initialize toolgroups/client_tools (#186) (458e207)
• broken .retrieve call using identifier= (#135) (626805a)
• bump to 0.2.1 (edb6173)
• bump version (b6d45b8)
• bump version in another place (7253433)
• cli: align cli toolgroups register to the new arguments (#231) (a87b6f7)
• correct toolgroups_id parameter name on unregister call (#235) (1be7904)
• fix duplicate model get help text (#188) (4bab07a)
• llama-stack-client providers list (#134) (930138a)
• react agent (#200) (b779979)
• React Agent for non-llama models (#174) (ee5dd2b)
• React agent should be able to work with provided config (#146) (08ab5df)
• react agent with custom tool parser n_iters (#184) (aaff961)
• remove the alpha suffix in run_benchmark.py (#179) (638f7f2)
• update CONTRIBUTING.md to point to uv instead of rye (3fbe0cd)
• update uv lock (cc072c8)
• validate endpoint url (#196) (6fa8095)

Chores

• api sync, deprecate allow_resume_turn + rename task_config->benchmark_config (Sync updates from stainless branch: yanxi0830/dev) (#176) (96749af)
• AsyncAgent should use ToolResponse instead of ToolResponseMessage ([#197](https:/...

v0.2.8

Release corresponding to Llama Stack v0.2.8

What's Changed

• fix: accept extra_headers in agent.create_turn and pass them faithfully by @ashwinb in #228
• feat: add client-side utility for getting OAuth tokens simply by @ashwinb in #230
• fix(cli): align cli toolgroups register to the new arguments by @dolfim-ibm in #231

New Contributors

• @dolfim-ibm made their first contribution in #231

Full Changelog: v0.2.7...v0.2.8

v0.2.0

Llama4 Support

Checkout more at https://www.llama.com/

Full Changelog: v0.1.9...v0.2.0

v0.1.9

Along with release in https://github.com/meta-llama/llama-stack/releases/tag/v0.1.9

Full Changelog: v0.1.8...v0.1.9

v0.1.8

Published along with https://github.com/meta-llama/llama-stack/releases/tag/v0.1.8

What's Changed

• feat: datasets api updates by @yanxi0830 in #203
• Sync updates from stainless branch: hardikjshah/dev by @hardikjshah in #204
• fix(agent): better error handling by @ehhuang in #207
• feat: add weighted_average aggregation function support by @SLR722 in #208
• fix: fix duplicate model get help text by @reidliu41 in #188
• simplify import paths; Sync updates from stainless branch: ehhuang/dev by @ehhuang in #205
• Sync updates from stainless branch: yanxi0830/dev by @yanxi0830 in #209

New Contributors

• @reidliu41 made their first contribution in #188

Full Changelog: v0.1.7...v0.1.8

v0.1.7

Published along with https://github.com/meta-llama/llama-stack/releases/tag/v0.1.7

What's Changed

• Add -h help flag support to all CLI commands by @alinaryan in #185
• feat: update react with new agent api by @yanxi0830 in #189
• feat: autogen llama-stack-client CLI reference doc by @yanxi0830 in #190
• chore: remove litellm type conversion by @ehhuang in #193
• chore: AsyncAgent should use ToolResponse instead of ToolResponseMessage by @ehhuang in #197
• fix: validate endpoint url by @cdoern in #196
• fix: react agent by @ehhuang in #200
• Sync updates from stainless branch: main by @yanxi0830 in #198
• chore: Sync updates from stainless branch: ehhuang/dev by @ehhuang in #199
• feat(agent): support multiple tool calls by @ehhuang in #192
• feat: llama-stack-client providers inspect PROVIDER_ID by @cdoern in #181
• chore: Sync updates from stainless branch: main by @yanxi0830 in #201
• fix: llama-stack-client provider inspect should use retrieve by @cdoern in #202

New Contributors

• @alinaryan made their first contribution in #185

Full Changelog: v0.1.6...v0.1.7

v0.1.6

v0.1.6

Published along with Llama Stack v0.1.6

What's Changed

• feat: unify max infer iters with server/client tools by @yanxi0830 in #173
• chore: api sync, deprecate allow_resume_turn + rename task_config->benchmark_config (Sync updates from stainless branch: yanxi0830/dev) by @yanxi0830 in #176
• fix: remove the alpha suffix in run_benchmark.py by @SLR722 in #179
• chore: Sync updates from stainless branch: ehhuang/dev by @ehhuang in #182
• feat: support client tool output metadata by @ehhuang in #180
• chore: use rich to format logs by @ehhuang in #177
• feat: new Agent API by @ehhuang in #178
• fix: react agent with custom tool parser n_iters by @yanxi0830 in #184
• fix(agent): initialize toolgroups/client_tools by @ehhuang in #186
• feat: async agent wrapper by @yanxi0830 in #169
• feat(agent): support plain function as client_tool by @ehhuang in #187

Full Changelog: v0.1.5...v0.1.6

v0.1.5

Published along with Llama Stack v0.1.5

What's Changed

• feat: add support for chat sessions by @cdoern in #167
• fix react agent by @hardikjshah in #172
• fix: React Agent for non-llama models by @hardikjshah in #174

Full Changelog: v0.1.4...v0.1.5

v0.1.4

What's Changed

• feat: Sync updates from stainless branch: ehhuang/dev by @ehhuang in #149
• Update CODEOWNERS by @SLR722 in #151
• chore: deprecate eval task (Sync updates from stainless branch: main) by @yanxi0830 in #150
• refine the benchmark eval UX by @SLR722 in #156
• fix: React agent should be able to work with provided config by @hardikjshah in #146
• feat (1/n): agents resume turn (Sync updates from stainless branch: yanxi0830/dev) by @yanxi0830 in #157
• v0.1.4 - Sync updates from stainless branch: yanxi0830/dev by @yanxi0830 in #164

New Contributors

• @hardikjshah made their first contribution in #146

Full Changelog: v0.1.3...v0.1.4