Get customers
Answer customers
Work for you
Reach people
Give agents tools
Run and ship
Manage prompt versions and releases, request feedback, tables and evaluations
List workspace AI workflows and retain provider pagination.
4 fields
Get raw chat or completion content and version metadata.
5 fields
Retrieve a template with input variables applied; does not run an LLM.
8 fields
Publish a new prompt or version, optionally moving release labels to it.
7 fields
Create a version from a base version or label, replacing selected fields.
List release labels for one prompt.
1 field
Create a release label pointing to a prompt version.
3 fields
Move an existing release label to a different version.
2 fields
Delete a release label without deleting its prompt.
Associate custom metadata with a logged request.
Track a named or default score from 0 to 100.
Associate a request with a prompt and its input variables.
Read a logged request's prompt, usage, timing and cost.
Read the spans belonging to a trace.
List tables in the key's workspace, retaining the next cursor.
6 fields
Create a table, optionally including a default sheet and text column.
Read a table and its sheet metadata.
List sheets belonging to one table.
Create a blank sheet with the provider's default text column scaffold.
Read sheet details and version count.
Read one page of rows with optional column and execution metadata.
9 fields
Append rows; computed cells remain stale until explicitly recalculated.
Queue recalculation; preserve any confirmation request instead of approving it automatically.
Poll a queued sheet recalculation without starting another run.
Read aggregate and per-column scores and calculation status.
