CompletionKit

Delete a dataset

Delete a metric group

Destroy a draft MetricVersion (use for either source: 'edit' or source: 'suggestion'). Published versions are refused. To demote a published version, publish a different one as current instead.

Delete a metric

Delete a prompt

Delete a provider credential

Delete a run

Delete a tag. Removes the tag from every linked metric, prompt, run, and dataset.

Publish a MetricVersion as the live version of its metric. Works for both 'draft → published' and 'revert to an older published version → current'. Transactionally flips current, demotes peers, and writes the version's instruction + rubric_bands back onto the metric so the judge grades against it.

Publish a prompt version, making it the current version

Upsert an agreement for (run, response, metric, created_by). Verdict is one of agree, disagree, borderline. corrected_score (1..5) is required when verdict is 'disagree'.

Create a dataset with CSV data. First row is the header. Two column names are recognized specially: "expected_output" is each row's answer key (ground truth) given to the judge and to checks that compare against the row's expected value, and "actual_output" is a pre-made output to score in a prompt-less run. Both are overridable per run (expected_column / output_column). Every column is also available to the prompt as a variable.

Create a dataset by downloading CSV from a URL instead of inlining it. Use this for large datasets: pass a public http(s) URL and the server fetches the CSV directly, so the data never has to pass through the tool-call arguments. The URL is SSRF-checked and the download is capped at 10MB. First row is the header; the "expected_output" (answer key) and "actual_output" (pre-made output) columns are recognized specially, overridable per run.

Update a dataset

Create a scoring run for the current judge over a dataset's existing outputs (wraps runs_create with prompt_id omitted and output_column supplied). This only sets up the run; call runs_generate to actually re-judge the outputs so you can compare against human verdicts.

Create a metric group

Update a metric group

Create a metric with evaluation criteria. For a deterministic check set metric_type:"check" and check_config. Per-kind required keys: value (contains/not_contains/equals/list_overlap/numeric_equals), pattern (regex), json_path+expected (json_path_equals), min and/or max (length_bounds/numeric_bounds); valid_json takes no extra keys. target_path is required when target is json_path; a target_path sent with no target sets target to json_path, and one sent with any other target is refused. target_path walks array indices, as in items.0.name; any path may start at the root with $., as in $.items.0.name. list_overlap grades a list-valued field and stores a fraction rather than a verdict: score_by is recall, precision, f1 or jaccard (recall by default), and min is the fraction that counts as a pass (1, a perfect score, by default). numeric_bounds and numeric_equals read a number, ignoring the thousands commas and currency symbol a figure is written with; a comma counts only as a thousands separator, with three digits after it and no decimal point before it, so 1,5 is not a number; numeric_equals takes a tolerance, either a plain number for a fixed amount or a percentage string such as 2% for a share of the expected value. A check with no case_sensitive matches case-insensitively, regex and list_overlap included; send case_sensitive: true to match case exactly. Before 0.29.0 a regex check read a missing case_sensitive as case-sensitive. Set compare_to:"expected" on contains, not_contains, equals, json_path_equals, list_overlap, numeric_equals to grade against each row's own expected_output (ground truth) instead of a constant; drop value (or expected, for json_path_equals) and add expected_path to dig into the answer key when it is JSON.

Ask the model to rewrite the metric's judge instruction in N variants targeted at the recent disagreements. Each variant is saved as a draft MetricVersion with source="suggestion". Returns the persisted drafts. Stripe-metering hooks fire via ActiveSupport::Notifications under completion_kit.judge_suggestion.generated.

Update a metric. For a deterministic check set metric_type:"check" and check_config. Per-kind required keys: value (contains/not_contains/equals/list_overlap/numeric_equals), pattern (regex), json_path+expected (json_path_equals), min and/or max (length_bounds/numeric_bounds); valid_json takes no extra keys. target_path is required when target is json_path; a target_path sent with no target sets target to json_path, and one sent with any other target is refused. target_path walks array indices, as in items.0.name; any path may start at the root with $., as in $.items.0.name. list_overlap grades a list-valued field and stores a fraction rather than a verdict: score_by is recall, precision, f1 or jaccard (recall by default), and min is the fraction that counts as a pass (1, a perfect score, by default). numeric_bounds and numeric_equals read a number, ignoring the thousands commas and currency symbol a figure is written with; a comma counts only as a thousands separator, with three digits after it and no decimal point before it, so 1,5 is not a number; numeric_equals takes a tolerance, either a plain number for a fixed amount or a percentage string such as 2% for a share of the expected value. A check with no case_sensitive matches case-insensitively, regex and list_overlap included; send case_sensitive: true to match case exactly. Before 0.29.0 a regex check read a missing case_sensitive as case-sensitive. Set compare_to:"expected" on contains, not_contains, equals, json_path_equals, list_overlap, numeric_equals to grade against each row's own expected_output (ground truth) instead of a constant; drop value (or expected, for json_path_equals) and add expected_path to dig into the answer key when it is JSON.

Import a promptfooconfig.yaml. Creates a prompt, a dataset from the test vars, and metrics from the assert blocks (llm-rubric/g-eval become judge metrics; contains/equals/regex/is-json become deterministic check metrics). Returns a summary of what mapped and what was skipped and why; nothing is dropped silently.

Create a prompt

Suggest an improved version of a prompt, grounded in a run's test results and judge feedback. Analyzes the run's responses, scores, and reviews, then returns reasoning plus a rewritten template (preserving {{variables}}) and persists it as a Suggestion. Requires a run that has a prompt (not a scoring-only run).

Update a prompt. If the prompt already has runs, this creates a new DRAFT version (current=false) rather than editing in place or publishing. Promote it with prompts_publish, so an agent's edits don't go live without a gate. If it has no runs, it is updated in place.

Create a provider credential

Update a provider credential. Changing provider or api_endpoint on a credential with a stored key requires api_key in the same call, so a stored key never moves to a new provider or endpoint.

Create a run. Omit prompt_id and provide output_column to score existing outputs by grading a pre-existing dataset column instead of generating new ones.

Start a run. Required for every run, including score-only runs (no prompt): generates responses with the prompt when there is one, otherwise copies the graded dataset column and grades it.

Get a run by ID, including "metric_averages": a per-metric breakdown with each metric's average score (or pass rate for checks), how many rows it graded, and how many scored low. Use this to find the metric dragging a prompt down without listing responses.

List all runs

Re-grade a run's existing responses with its currently attached metrics, without regenerating. Use after attaching or editing metrics on an already-generated run.

Create and start a fresh copy of a run with the same prompt, dataset, metrics, and settings. Use when the judge changed and you want a clean run instead of mixing versions.

Re-run only the failed responses of a run, optionally limited to specific response ids via "only".

Update a run

Create a tag. Color is auto-assigned.

Rename a tag.

List agreements. Filter by run_id, response_id, metric_id, or created_by.

Get a dataset by ID

List all datasets

Compare two versions of one metric's agreement stats side by side. Requires metric_id, metric_version_a_id, and metric_version_b_id (both versions must belong to that metric). Unavailable for check metrics.

Get a metric group by ID

List all metric groups

List every MetricVersion (drafts + published) for a metric, newest first. Each row carries version_number, state, source, current flag, and timestamps.

Get a metric by ID

List all metrics

Get a prompt by ID

List all prompts

Get a provider credential by ID (API key is not exposed)

List all provider credentials (API keys are not exposed)

Get a specific response

List responses for a run, in row order. Returns {total, limit, offset, returned, responses}. Defaults to 50 rows because full payloads are large: use "fields" to drop the bodies, "min_score"/"max_score" to isolate low scorers, and sort "score_asc" to read the worst rows first. For per-metric averages of the whole run use runs_get instead of aggregating here.

Get a tag by ID

List all tags

Get this organization's plan usage and limits for the current billing period: runs and prompt fetches used, their limits, how many remain, and when the period resets. Call this to pre-check quota before starting runs. Runs are hard-blocked once the run limit is reached (with a small grace band), so a run over the limit will fail with run_limit_reached.