API Reference

View the Promptic API schema. Download OpenAPI spec.

AI Application

Get current AI application GET

Retrieve details for the authenticated AI application.

List AI applications GET

List all AI applications for the current user.

Components

List all components GET

Retrieve all AI components in your AI application.

Create a component POST

Create a new AI component to organize prompts and traces.

Get a component GET

Retrieve details for a specific component.

Experiments

Create an experiment POST

Set up a new prompt optimization experiment.

Start an experiment POST

Start a pending experiment to begin optimization.

Duplicate an experiment POST

Clone an experiment with its observations and evaluators; optionally seed from the source's best optimized prompt.

Get best iteration GET

Retrieve the highest-scoring iteration.

Evaluators & Observations

Create evaluator(s) POST

Add evaluation criteria to score prompt performance.

Create observation(s) POST

Add expected outputs and input variables for training and evaluation.

Deployment

Deploy experiment POST

Deploy an optimized experiment to a component.

Get deployed prompt GET

Retrieve the currently deployed prompt, model, and score.

Traces

List traces GET

Query traces with filtering by status, component, and time range.

Ingest OTLP traces POST

Send OpenTelemetry trace data in JSON or protobuf format.

Get tracing stats GET

Get aggregated tracing statistics for your AI application.

Storage

Create a presigned upload POST

Reserve a storage object and get a short-lived direct-upload descriptor.

Get a download URL GET

Return a short-lived signed GET URL for a storage object.

Artifacts

List artifacts for a trace GET

List the artifacts referenced by a trace's spans.

Upload an artifact POST

Get artifact metadata GET

Retrieve metadata and a content URL for an artifact.

Download artifact content GET

Redirect to a short-lived provider URL for the artifact bytes.

Datasets & Runs

Create a dataset POST

Create a dataset to group traces for evaluation.

Create a run POST

Create an evaluation run against a dataset.

Evaluations

Create an evaluation POST

Trigger an evaluation against a dataset or run.

List evaluations GET

List evaluations for a component, scoped by target (trace, run, or dataset).

Get an evaluation GET

Retrieve an evaluation with its current status, target anchor, and legacy results blob.

List judge results GET

List canonical per-target judge rows (score, rating, rationale, evidence, judgeSnapshot, version) for an evaluation.