Runs
List a run step's agent conversation
List the agent conversation of a run step. Requires permission runs:read.
GET
List the agent conversation of a run step
The reference panel describes the current v3 operation. The notes below describe its v2 compatibility API. Follow the v3 request schema for current calls: v3 can move identifiers from the body or query into the URL. Keep the v2 method, path, and payload together when using the compatibility API.
v2 compatibility notes
Returns one page of one step’s agent conversation, oldest first.stepId is a step id GET /api/v2/runs/getTimeline returns. A step that ran no agent answers an empty last page; a step the run does not have answers 404 “Step not found”.
Each message has its id, its order (a turn’s messages share one) and stepOrder within the turn, its role (system, user or assistant; tool results arrive inside the assistant message’s toolCalls), its status, its text, the agentName that wrote it, its reasoning, and its toolCalls (each a toolName, its input, its output or error text, and the AI SDK state). Credential-shaped strings (JWTs, Bearer tokens, x-access-token: URLs and GitHub tokens) are replaced by a [REDACTED] marker. Never the agent thread, who wrote the message, or the message’s raw parts or metadata.
Paging
limit is 1 to 50 (default 20). A turn’s messages share an order, so one turn may span two pages. A page may be shorter than limit while meta.hasMore is true; follow meta.nextCursor until it is false. A cursor the server cannot read answers 400.
Visibility
You needruns:read at the run’s home (its own org unit, else its workflow’s owning org unit) and at every company it carries, the same rule as GET /api/v2/runs/get. Any other run, and a run with neither home, answers 404 “Run not found” like a missing one. A Personal key reads as its owner, cut down to the key’s scopes; an External App reads by its own role grants.Authorizations
Query Parameters
Response
One page of the step's messages

