NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
PyPI, #308 most downloaded on PyPI
Client library to connect to the LangSmith Observability and Evaluation Platform.
Latest stable
0.16.0
Last release today
21 Nov 2026
Ships on a steady schedule
a new release about every 1 weeks
Nearly every release is documented
notes for 60 of the last 60 stable releases
9 versions withdrawn
withdrawn after publishing
3 years old
553 releases · first in 2023
Drag across the ruler to compare, or tap a release to open it.
feat(js): Use new multipart endpoint when available by @jacoblee93 in https://github.com/langchain-ai/langsmith-sdk/pull/1106
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.137...v0.1.138
Your coding agent can read these notes before it upgrades. Set up the MCP server →
integration tests against staging by @efriis in https://github.com/langchain-ai/langsmith-sdk/pull/1143
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.138rc1...v0.1.138rc2
feat(js): Use new multipart endpoint when available by @jacoblee93 in https://github.com/langchain-ai/langsmith-sdk/pull/1106
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.137...v0.1.138rc1
fix: Create new multipart encoder for each retry attempt by @nfcampos in https://github.com/langchain-ai/langsmith-sdk/pull/1114
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.136...v0.1.137
feat(JS): log wrapOpenAI calls with usage metdata by @agola11 in https://github.com/langchain-ai/langsmith-sdk/pull/1096
wrapOpenAI calls with usage metdata by @agola11 in https://github.com/langchain-ai/langsmith-sdk/pull/1096processInputs / processOutputs to traceable + add usage metadata to wrapOpenAI by @agola11 in https://github.com/langchain-ai/langsmith-sdk/pull/1095Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.135...v0.1.136
feat(js): Validate that store values are RunTree instances by @jacoblee93 in https://github.com/langchain-ai/langsmith-sdk/pull/1088
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.134...v0.1.135
py: Add session_id and session_name to patch parts in batch/multipart by @nfcampos in https://github.com/langchain-ai/langsmith-sdk/pull/1086
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.133...v0.1.134
Update README.md by @eyurtsev in https://github.com/langchain-ai/langsmith-sdk/pull/1072
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.132...v0.1.133
py: For multipart endpoint don't split into batches of 20mb by @nfcampos in https://github.com/langchain-ai/langsmith-sdk/pull/1067
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.131...v0.1.132
Run baseline benchmark job whenever bench code changes by @nfcampos in https://github.com/langchain-ai/langsmith-sdk/pull/1058
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.130...v0.1.131
chore: bump sdk 0.7.39 by @langchain-infra in https://github.com/langchain-ai/langsmith-sdk/pull/1047
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.129...v0.1.130
[JS] Catch any errors in create/update run by @hinthornw in https://github.com/langchain-ai/langsmith-sdk/pull/1044
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.128...v0.1.129
[Python] increase default client timeout in evals by @hinthornw in https://github.com/langchain-ai/langsmith-sdk/pull/1030
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.127...v0.1.128
fix(js): Trim trailing slashes on passed API and web URLs by @jacoblee93 in https://github.com/langchain-ai/langsmith-sdk/pull/1033
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.126...v0.1.127
Adding method for removing a run from an annotation queue (#1027) by @hinthornw in https://github.com/langchain-ai/langsmith-sdk/pull/1029
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.125...v0.1.126
[Py] Default to unnamed by @hinthornw in https://github.com/langchain-ai/langsmith-sdk/pull/1024
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.124...v0.1.125
Change the 'text' model type to 'llm' to match other libraries by @efriis in https://github.com/langchain-ai/langsmith-sdk/pull/1019
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.123...v0.1.124
Fix two descriptions in python sdk by @nhuang-lc in https://github.com/langchain-ai/langsmith-sdk/pull/1016
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.122...v0.1.123
fix(js): Removes langchain and @langchain/core from peer deps to fix npm installation by @jacoblee93 in https://github.com/langchain-ai/langsmith-sdk/
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.121...v0.1.122
[Python] Use _name by @hinthornw in https://github.com/langchain-ai/langsmith-sdk/pull/1008
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.120...v0.1.121
Support evaluate on existing experiment by @hinthornw in https://github.com/langchain-ai/langsmith-sdk/pull/1000
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.119...v0.1.120
fix(js): Fix overzealous circular reference detection by @jacoblee93 in https://github.com/langchain-ai/langsmith-sdk/pull/998
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.118...v0.1.119
feat(js): Allow overriding fetch implementation used by LangSmith by @jacoblee93 in https://github.com/langchain-ai/langsmith-sdk/pull/989
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.117...v0.1.118
[Py] Lazy client initialization in run tree by @hinthornw in https://github.com/langchain-ai/langsmith-sdk/pull/983
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.116...v0.1.117
Exclude non-prompt|llm manifests by @hinthornw in https://github.com/langchain-ai/langsmith-sdk/pull/980
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.115...v0.1.116
Nothing published for this version
[Python] Evals: submit feedback in thread by @hinthornw in https://github.com/langchain-ai/langsmith-sdk/pull/977
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.114...v0.1.115
Nothing published for this version
Nothing published for this version
feat: metadata filters for few shot by @jakerachleff in https://github.com/langchain-ai/langsmith-sdk/pull/974
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.113...v0.1.114
[Python] Log content-length during connection errors by @hinthornw in https://github.com/langchain-ai/langsmith-sdk/pull/972
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.112...v0.1.113
[Python] Cache env var check by @hinthornw in https://github.com/langchain-ai/langsmith-sdk/pull/971
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.111...v0.1.112
Ensure if null by @hinthornw in https://github.com/langchain-ai/langsmith-sdk/pull/969
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.110...v0.1.111
[Python] Async create feedback from token by @hinthornw in https://github.com/langchain-ai/langsmith-sdk/pull/960
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.109...v0.1.110
Filter child runs where trace has been filtered by @hinthornw in https://github.com/langchain-ai/langsmith-sdk/pull/964
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.108...v0.1.109
Filter Root (not children) by @hinthornw in https://github.com/langchain-ai/langsmith-sdk/pull/961
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.107...v0.1.108
Nothing published for this version
Optimize CPU-bound Task Execution by Replacing Thread Lock with Producer-Consumer Pattern by @adham-elarabawy in https://github.com/langchain-ai/langs
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.106...v0.1.107
[JS] Fix summary evaluator result logging by @hinthornw in https://github.com/langchain-ai/langsmith-sdk/pull/955
Full Changelog: https://github.com/langchain-ai/langsmith-sdk/compare/v0.1.105...v0.1.106
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
One column per quarter.