Output Handling — Fluxara Docs

Artifact model

An artifact is a file produced by a completed job. In the current beta scope, artifacts are typically image outputs accompanied by metadata describing the pipeline, preset, and execution context.

Artifacts are intended to be easy to retrieve, review, compare, and export across internal workflows.

Typical artifact fields

Artifact ID

A unique identifier used to reference the output in the current workspace or job context.

Type

The file type of the output, such as image/png or image/webp.

Dimensions

Width and height values for visual artifacts generated by the selected workflow.

Creation timestamp

The time at which the artifact was finalized and made available for retrieval.

Job reference

A link back to the job that produced the artifact and the pipeline context used during execution.

Download path

A retrieval URL or artifact endpoint exposed to approved beta users in supported environments.

Associated metadata

Output artifacts are typically accompanied by metadata that helps teams understand how a result was produced and how it should be compared against other runs.

prompt or input summary
selected preset
pipeline identifier
resolution or size profile
inference step count
artifact format
job creation and completion timestamps

Output lifecycle

Job Created Processed Artifact Written Metadata Recorded Available for Retrieval

This lifecycle helps separate execution state from output state. A job may be running or queued without any available artifact, while completed jobs expose artifacts and metadata together.

Export formats

The current beta focuses on common image output formats and structured metadata suitable for internal review workflows.

PNG

Used where predictable lossless output is preferred for review, inspection, or archival comparison.

WebP

Used where smaller output size is preferred while preserving practical visual quality for browser-based workflows.

Structured metadata

Associated metadata may be returned alongside outputs through the API or internal retrieval interfaces.

Preview-friendly delivery

Outputs are designed to work well in lightweight review and comparison flows during the beta period.

Retention behavior

Artifact retention may vary depending on environment capacity, onboarding scope, and workspace policy during the beta period.

In practice, retention should be treated as operational rather than archival. Teams evaluating the platform should not assume indefinite artifact storage unless explicitly agreed during onboarding.

retention windows may differ across environments
temporary artifact access may be used in some workflows
older outputs may be removed as storage policies evolve
metadata availability may outlast binary artifact availability

Review workflows

Output handling is designed to support internal review rather than public publishing workflows. A typical review flow includes:

Generate variants

Produce a comparable set of outputs using a shared preset or pipeline structure.

Inspect metadata

Confirm that outputs were created under the intended execution conditions.

Select preferred result

Narrow multiple candidates down to the artifact best suited to the target workflow.

Export or archive locally

Move selected artifacts into local storage or downstream internal tooling when needed.

Current beta constraints

Retention behavior may change as storage policies mature.
Not all outputs are guaranteed to remain available indefinitely.
Export options may differ by environment.
Artifact access is limited to approved beta contexts.

Design principles

Traceability

Every artifact should be understandable in the context of the job and preset that produced it.

Practical retrieval

Outputs should be easy to access and inspect without adding unnecessary workflow complexity.

Review-first design

Output handling is optimized for internal evaluation and selection workflows.

Operational flexibility

Retention and delivery behavior can evolve during the beta period without breaking the overall artifact model.