langsmith-docs

mirror of https://github.com/langchain-ai/langsmith-docs.git synced 2026-07-16 01:12:20 -04:00

Files

T

History

Paarth Ahuja 1edf4e5a03 docs(eval): inline pairwise judging prompt instead of pulling from hub (#1002 )

The "Run a pairwise evaluation" example calls
`langchain.hub.pull("langchain-ai/pairwise-evaluation-2")`, which raises
`ValueError` on current `langsmith` versions: pulling public prompts now
requires `dangerously_pull_public_prompt=True`, and neither
`langchain.hub.pull` nor `langchain_classic.hub.pull` forwards the flag.

Two fixes were considered: (a) call `Client().pull_prompt(...,
dangerously_pull_public_prompt=True)` directly and add a callout, or (b)
inline the prompt and schema. This PR takes (b) — the system and human
messages are copied verbatim and `Score` is declared as a Pydantic model
bound via `with_structured_output`. It removes the dependency on the
broken wrapper, makes the schema visible on the page, and matches the
TypeScript example on the same page.

2026-05-05 11:34:49 -04:00

administration

docs: update api key docs for org-scoped keys (#959 )

2025-09-03 15:34:21 -04:00

evaluation

docs(eval): inline pairwise judging prompt instead of pulling from hub (#1002 )

2026-05-05 11:34:49 -04:00

observability

docs: Adds ttft run tree docs (#943 )

2025-08-13 11:43:02 -07:00

pricing

Update sales links (#950 )