Skip to main content
GET
Rate limit tier: general, 600 requests/min. See Rate limits. Fetches one eval by its eval_id. It is the poll target for a run launched with POST /v3/workflows/{workflow_id}/evals/. The id in the response is the identifier that launch returned as eval_id. The response is the eval’s frozen headline: grading status, the accuracy fraction, the matched, mismatched and ungraded field tallies, the cohort’s file_count, and failed_count. failed_count counts the documents that failed to enqueue. The headline carries counts, never internal ids.

Status model

The endpoint always returns 200 while the eval exists. Grading in flight raises no precondition error. Poll until status is terminal. On a processed eval, accuracy is the matched / (matched + mismatched) fraction. It stays null when the graded denominator is 0, which happens when nothing is gradable. Every field being ungraded for lack of ground truth produces that case. An unknown id returns 404, and so does an eval on another workflow or organization: existence never leaks. A caller outside the workflow’s organization gets 403.

Authorizations

Authorization
string
header
required

API key issued from app.anyformat.ai/api-key. Send as Authorization: Bearer <key>.

Path Parameters

workflow_id
string
required
eval_id
string
required

Response

Successful Response

One eval's frozen headline — the poll target for a launched run.

id here is the same identifier the launch returned as eval_id. The headline fields (accuracy, matched, mismatched, ungraded) are null while status is in_progress; accuracy stays null on a processed eval whose graded denominator is 0.

id
string
required

Unique identifier of the eval (hyphenated UUID); equals the launch's eval_id.

Example:

"069dcc2c-e14c-7606-8000-2ee4fb17b4f9"

run_number
integer
required

Per-workflow run counter (#1, #2, …) for this eval.

Example:

3

version_id
string
required

Workflow version the eval scored.

Example:

"abcdef1234"

status
enum<string>
required

Grading status. Non-terminal: in_progress. Terminal: processed, error.

Available options:
in_progress,
processed,
error
Example:

"processed"

file_count
integer
required

Number of dataset documents in the eval's realized cohort.

Example:

42

failed_count
integer
required

Number of dataset documents that failed to enqueue for this run.

Example:

0

accuracy
number | null

Headline accuracy in [0, 1]; null while in_progress and on a processed eval with a 0 graded denominator.

Example:

0.94

matched
integer | null

Fields that matched ground truth; null while in_progress.

mismatched
integer | null

Fields that mismatched ground truth; null while in_progress.

ungraded
integer | null

Fields with no ground truth to grade against; null while in_progress.

created_at
string<date-time> | null

Timestamp when the eval was created (ISO 8601).