Skip to main content
GET
Get evaluation result

Authorizations

Authorization
string
header
required

SDK API key (tl_...) or Clerk session JWT

Path Parameters

id
string<uuid>
required

Response

Evaluation

id
string<uuid>
required
trace_id
string<uuid>
required
project_id
string<uuid>
required
metric_name
string
required
evaluator_type
enum<string>
required
Available options:
llm,
rule
status
enum<string>
required
Available options:
pending,
running,
complete,
failed
context
enum<string>
required

Distinguishes a fresh evaluation against a newly-ingested trace (fresh) from a re-score the eval runner stamped after the executor re-published a remediated trace (remediation). Defaults to fresh for evaluations triggered before remediation tracking shipped (TRU-370).

Available options:
fresh,
remediation
score
number | null
label
string | null
reasoning
string | null
source_action_id
string<uuid> | null

Control action that drove this remediation re-score (only set when context = remediation). Points at GET /v1/control/actions/{id} so the dashboard can show "this eval was triggered by THAT action".

original_evaluation_id
string<uuid> | null

Evaluation whose score drove the policy decision behind source_action_id — the baseline the regression detector compares this eval against. Null on fresh evaluations.